Syed is currently unavailable
Syed Ali Azzam
Verified Expert in Engineering
AI Consultant and Developer
Karachi, Sindh, Pakistan
Toptal member since February 11, 2026
Bio
As Founder and CEO of Paramount Intelligence, Ali helps Fortune 1,000 teams and fast-growing companies accelerate ROI from AI by designing and engineering systems that run inside existing workflows. He combines applied AI delivery with strong data foundations, analytics, and automation so solutions move beyond pilots and become reliable day-to-day tools. Ali works hands-on with client leaders to identify high-value opportunities and deliver measurable operational outcomes with speed and quality.
Portfolio
Paramount Intelligence
AI Systems, Data Analytics, Data Engineering, Data Visualization, Data Modeling...
Gratia
AI Systems, Data Analytics, Data Modeling, Data Visualization...
Syngenta
GitLab CI/CD, AWS Cloud Development Kit (CDK), JWT Authentication...
Experience
- Retrieval-augmented Generation (RAG) - 5 years
- LangChain - 5 years
- AWS (Lambda + API Gateway) - 5 years
- Python - 5 years
- Data Engineering - 5 years
- AI Systems - 5 years
- Bedrock - 4 years
- AWS Bedrock AgentCore - 3 years
Preferred Environment
Git, LangChain, PostgreSQL, Google Workspace, Python, AWS (Lambda + API Gateway), FastAPI, Docker, Visual Studio Code (VS Code), n8n
The most amazing...
...thing I’m proudest of is shipping an AI and automation system that saves a Fortune 1,000 team millions of dollars per month in costs.
Work Experience
Founder & CEO
2025 - PRESENT
Paramount Intelligence
- Led AI strategy, architecture, and enterprise delivery for Fortune 1,000 and high-growth companies, owning engagements end-to-end from problem definition to production deployment.
- Architected and deployed AI chatbots, knowledge systems, and workflow automation that replaced manual operations and saved millions of dollars in annual operating costs.
- Partnered directly with founders and C-suite executives to translate business goals into scalable AI systems, improving margins, accelerating operations, and unlocking new revenue.
Technologies: AI Systems, Data Analytics, Data Engineering, Data Visualization, Data Modeling, Decision Analytics, Business Intelligence (BI), AWS (Lambda + API Gateway), ETL Pipelines, Google Workspace, FastAPI, Experimental Design, KPI & Performance Analysis, LLM Applications, Machine Learning, PostgreSQL, n8n, LangChain, Python for Analytics, Predictive Analytics, Statistical Modeling, Python, SQL, Visual Studio Code (VS Code), Workflow Automation, Git, Docker, Large Language Models (LLMs), Artificial Intelligence (AI), Deep Learning, Real-time Systems, Spatial Analysis, Combinatorial Optimization, Computer Vision, Payment APIs, Payments, Data Science, IT Management, Azure, Claude API, Anthropic, Claude, AI Agents, Agentic AI, APIs, React, TypeScript, AI Copilots, ChatGPT, ChatGPT API, ChatGPT Prompts, Claude Agent SDK, IT Project Management, IT Projects, Project Management, Metal, OpenCV, C++, NVIDIA Jetson, You Only Look Once (YOLO), Edge Computing, Radio Frequency (RF) Protocol, Thermal Imaging, Natural Language Processing (NLP), Prompt Engineering, Android, App Intents, CI/CD Pipelines, Expo, GitHub, Gmail API, Google Calendar, LangGraph, Mobile Applications, OAuth, OpenAI API, Push Notifications, REST APIs, React Native, Speech-to-Text (STT), Machine Learning Operations (MLOps), Image Processing, Neural Networks, Google Cloud Platform (GCP), AI Programming, AI Tools, Airtable, JavaScript, Vapi, AI Voice Agents, Solution Architecture, Agentic Frameworks, Architecture, Large Language Model Operations (LLMOps), Lovable, RAG Pipelines, Microsoft Copilot, AI Security, Generative Artificial Intelligence (GenAI), Reinforcement Learning from Human Feedback (RLHF), PyTorch, XGBoost, Cybersecurity, Networking, AI-generated Video, Workflows, Data Privacy, Identity & Access Management (IAM), Personally Identifiable Information (PII), Vector Databases, LlamaIndex, Analytics, Snowflake, Tableau, Financial Data, Model Context Protocol (MCP), Pydantic, MongoDB, API Development, Generative Pre-trained Transformers (GPT), OpenAI, AWS Deployment, AWS DevOps, Amazon Web Services (AWS), Automation, DevOps, Infrastructure, AI Design, Gemini Enterprise, Vertex AI, JSON, Natural Language Toolkit (NLTK), Data Extraction, Data Pipelines, Data Labeling, Databricks, Gensim, Information Extraction, PySpark, SpaCy, AI Architecture, AI Consulting, AI Integration, Cloud Architecture, Microsoft 365, RAG Architecture, ERP Systems, HubSpot CRM, Odoo, Full-stack Development, Image Generation, Forecasting, System Modeling, Retrieval-augmented Generation (RAG), Agentic RAG Systems, Data Analysis, Attribution Modeling, Customer Segmentation, Funnel Marketing, Markov Model, Marketing, Marketing Attribution, Natural Language Queries, Natural Language Understanding (NLU), A/B Testing, Model Validation, ARIMA, NumPy, Pandas, BigQuery, StatsModels, Multi-agent Systems, System Design, Supabase, Stripe, Regulatory Compliance, Data Architecture, Business Rules Engine, Parse, NoSQL, Event-driven Architecture, RAG Systems, Security, Technical Leadership, Data Quality, Transformers, Apache Parquet, Parquet, Scikit-learn, Seaborn, Prefect, Hetzner, Quantitative Analysis, CrewAI, AWS Fargate, AWS IAM, AWS Global Accelerator, AIOps, Claude Code, Cursor AI, Healthcare, Jinja, Optical Character Recognition (OCR), PDF, Azure AI Document Intelligence, Engines, ML Pipelines, Scalability, Semantic Search, Qdrant, FAISS, Weaviate, Front-end Development, Document Processing, Data Governance, Legal, AWS Lambda, Amazon S3 (AWS S3), Cost Reduction & Optimization (Cost-down), Document Parsing, Model Evaluation, Source Code Review, System Architecture, AI Chatbots, Amazon Bedrock, ETL, Jupyter Notebook, Recommendation Systems, Risk Modeling, Risk Models, Web Scraping, Data-informed Recommendations, ETL Tools, Real-time Data, Streaming Data, Website Data Scraping, Full-stack, HIPAA Compliance, Kubernetes, Custom Models, Pinecone, Finance, Algorithms, Microsoft Copilot Studio, RESTFul APIs, Semantic Kernel (SK), Terraform, Foundry, Low Code, Azure AI Studio, Azure OpenAI Service, C#, GraphQL API, Bicep, Microsoft Power Platform, Azure AI Search, AutoGen, Voice Chat, WhatsApp Business, WhatsApp API, Data Build Tool (dbt), Data Integration, AWS Database Migration Service (DMS), AWS Glue, Cortex, Financial Analysis, Reporting, Distributed Systems, Back-end, Cloud Platforms, Flask, Workflow Automation & System Integration, CRM APIs, Debugging, Platform Design, API Design, AI SDK, AI Model Training, Training, MLflow, API Integration, Scraping, Real Estate, WhatsApp, Rule-based Programming, Benchmarking, Hyperparameter Tuning, OpenAI GPT-4 API, Automation Tools, Data Validation, Automations, Calendly API, DocuSign, Email Integration, Zoom, Asynchronous Programming, Webhooks, AI Automation, Consulting, Amazon QuickSight, Dagster, Inventory Management, Invoice Management, Leadership, SDKs, Team Leadership, Databases, Sales Forecasting, Vector Search, Machine Learning (ML) APIs, Chatbot Conversation Design, Chatbots, OpenAI SDK, Platform Engineering, Vercel, Agentic AI Systems, Governance, Regression Testing, Unity Catalog, Decision Modeling, Decision Trees, Enterprise Architecture, Stakeholder Management, Zapier, AI Agent Orchestration, Harness, Light LLMs, Pgvector, MySQL, Data Cleaning, Google Street View, Linear Regression, Statistical Analysis, Statistical Learning, CRM, Perplexity, Gemini, Whisper, Audio, Facial Recognition, Facial Tracking, CCTV, Video Surveillance, DeepFace, AdaFace, ArcFace, CosFace, Dlib, FaceNet, FaceNet512, GhostFaceNets, InsightFace, MagFace, Bayesian Statistics, Pytest, Statistics, Streamlit, Datadog, Coding, Google BigQuery, Patreon, Fintech, OpenAI Assistants API, Azure ML Studio, Azure Machine Learning, Markdown, Hooks, AWS Bedrock AgentCore
Senior AI Consultant (Independent Contractor)
2025 - PRESENT
Gratia
- Designed and maintained AI-driven operational systems and workflow automation, eliminating manual administrative work and delivering $150,000+ in annual productivity savings.
- Owned end-to-end API integrations between web forms, CRMs, ad platforms, and internal tools, ensuring reliable data flow and eliminating duplicate effort, saving $80,000+ annually in manual labor costs.
- Acted as an Operations Automation Expert, diagnosing bottlenecks and deploying workflows that cut cycle times by 30-50% and unlocked $200,000+ in annual capacity across teams.
- Built AI agents, conversational funnels, and an AI-driven applicant tracking system (ATS) assessment method that improved hiring speed by 80% and generated $100,000+ in incremental yearly revenue.
Technologies: AI Systems, Data Analytics, Data Modeling, Data Visualization, Decision Analytics, Data Engineering, Business Intelligence (BI), FastAPI, ETL Pipelines, Google Workspace, KPI & Performance Analysis, LangChain, LLM Applications, Machine Learning, n8n, PostgreSQL, Predictive Analytics, Python, Python for Analytics, SQL, Statistical Modeling, Visual Studio Code (VS Code), Workflow Automation, AWS (Lambda + API Gateway), Large Language Models (LLMs), Artificial Intelligence (AI), Deep Learning, GPS, GPS Tracker, Real-time Systems, Spatial Analysis, Combinatorial Optimization, Computer Vision, Payment APIs, Payments, Data Science, IT Management, Azure, Claude API, Anthropic, Claude, AI Agents, Agentic AI, APIs, TypeScript, AI Copilots, ChatGPT, ChatGPT API, ChatGPT Prompts, Claude Agent SDK, IT Project Management, IT Projects, Project Management, OpenCV, C++, NVIDIA Jetson, You Only Look Once (YOLO), Edge Computing, Radio Frequency (RF) Protocol, Thermal Imaging, Natural Language Processing (NLP), Prompt Engineering, Android, App Intents, CI/CD Pipelines, Expo, GitHub, Gmail API, Google Calendar, LangGraph, Mobile Applications, OAuth, OpenAI API, REST APIs, React Native, Speech-to-Text (STT), Machine Learning Operations (MLOps), Neural Networks, Google Cloud Platform (GCP), AI Programming, AI Tools, Airtable, JavaScript, Vapi, AI Voice Agents, Solution Architecture, Agentic Frameworks, Architecture, Large Language Model Operations (LLMOps), Lovable, RAG Pipelines, Microsoft Copilot, Generative Artificial Intelligence (GenAI), Reinforcement Learning from Human Feedback (RLHF), PyTorch, XGBoost, Networking, AI-generated Video, Workflows, Identity & Access Management (IAM), Personally Identifiable Information (PII), Vector Databases, LlamaIndex, Analytics, Snowflake, Model Context Protocol (MCP), Pydantic, MongoDB, API Development, Generative Pre-trained Transformers (GPT), OpenAI, AWS Deployment, AWS DevOps, Amazon Web Services (AWS), Automation, DevOps, Infrastructure, AI Design, Gemini Enterprise, Vertex AI, JSON, Natural Language Toolkit (NLTK), Data Extraction, Data Pipelines, Data Labeling, Databricks, Gensim, Information Extraction, PySpark, SpaCy, AI Architecture, AI Consulting, AI Integration, Cloud Architecture, Microsoft 365, RAG Architecture, ERP Systems, HubSpot CRM, Odoo, Forecasting, System Modeling, Retrieval-augmented Generation (RAG), Agentic RAG Systems, Data Analysis, Attribution Modeling, Customer Segmentation, Funnel Marketing, Markov Model, Marketing, Marketing Attribution, Natural Language Queries, Natural Language Understanding (NLU), A/B Testing, Model Validation, ARIMA, NumPy, Pandas, BigQuery, StatsModels, Multi-agent Systems, System Design, Supabase, Stripe, Regulatory Compliance, Data Architecture, Business Rules Engine, Parse, NoSQL, Docker, Event-driven Architecture, RAG Systems, Security, Technical Leadership, Data Quality, Transformers, Apache Parquet, Parquet, Scikit-learn, Seaborn, Prefect, Hetzner, Quantitative Analysis, CrewAI, AWS Fargate, AWS IAM, AWS Global Accelerator, AIOps, Claude Code, Cursor AI, Jinja, Optical Character Recognition (OCR), PDF, Engines, ML Pipelines, Scalability, Semantic Search, Qdrant, FAISS, Weaviate, Front-end Development, Document Processing, Data Governance, Legal, AWS Lambda, Amazon S3 (AWS S3), Cost Reduction & Optimization (Cost-down), Document Parsing, Model Evaluation, Source Code Review, System Architecture, AI Chatbots, Amazon Bedrock, ETL, Jupyter Notebook, Recommendation Systems, Risk Modeling, Risk Models, Web Scraping, Data-informed Recommendations, ETL Tools, Real-time Data, Streaming Data, Website Data Scraping, Full-stack, Kubernetes, Custom Models, Pinecone, Finance, Algorithms, Microsoft Copilot Studio, Low Code, Azure AI Studio, Azure OpenAI Service, C#, GraphQL API, Bicep, Microsoft Power Platform, Azure AI Search, AutoGen, Voice Chat, Data Build Tool (dbt), Data Integration, AWS Database Migration Service (DMS), AWS Glue, Cortex, Distributed Systems, Back-end, Cloud Platforms, Flask, Workflow Automation & System Integration, CRM APIs, Debugging, Platform Design, API Design, AI SDK, AI Model Training, Training, MLflow, API Integration, Scraping, WhatsApp, Rule-based Programming, Benchmarking, Hyperparameter Tuning, OpenAI GPT-4 API, Automation Tools, Data Validation, Automations, Calendly API, DocuSign, Email Integration, Zoom, Asynchronous Programming, Webhooks, AI Automation, Consulting, Amazon QuickSight, Dagster, Inventory Management, Invoice Management, Leadership, SDKs, Team Leadership, Databases, Sales Forecasting, Vector Search, Machine Learning (ML) APIs, Chatbot Conversation Design, Chatbots, Platform Engineering, Vercel, Agentic AI Systems, Governance, Regression Testing, Unity Catalog, Decision Modeling, Decision Trees, Enterprise Architecture, Stakeholder Management, Zapier, AI Agent Orchestration, Harness, Light LLMs, Pgvector, MySQL, Data Cleaning, Linear Regression, CRM, Perplexity, Gemini, Whisper, Audio, Coding, Patreon, Fintech, OpenAI Assistants API, Azure ML Studio, Azure Machine Learning, Markdown, Hooks, AWS Bedrock AgentCore
AI Platform Engineer | AWS Bedrock & AgentCore
2026 - 2026
Syngenta
- Designed and delivered the Syngenta AI Foundry AgentCore Registry, a centralized AI governance platform for discovering, managing, and governing AI agents and MCP (Model Context Protocol) gateways across enterprise AWS environments.
- Architected and implemented a serverless AWS Bedrock AgentCore solution using AWS Lambda, API Gateway, Amazon S3, and CloudFront to support AI asset discovery, metadata management, and lifecycle governance.
- Built automated ingestion pipelines that discover and catalog AI agents, AgentCore runtimes, and MCP gateways, enabling centralized visibility across distributed enterprise AI ecosystems.
- Developed a complete governance workflow supporting Draft, Submit, Pending Approval, Approved, Rejected, and Deprecated states, establishing enterprise-grade controls for AI asset lifecycle management.
- Implemented role-based access control using AWS Cognito, JWT authentication, custom authorization logic, and persona-based permissions for consumers, publishers, and administrators.
- Engineered scalable backend APIs using Python, Boto3, REST APIs, and AWS Lambda, providing catalog management, governance operations, metadata updates, search, filtering, and approval workflows.
- Designed metadata normalization and registry services capable of supporting A2A agents, MCP gateways, custom AI assets, and heterogeneous agent architectures through a unified governance model.
- Automated deployment workflows using GitLab CI/CD, OIDC authentication, Terraform, Serverless Framework, and Infrastructure as Code practices, eliminating static cloud credentials and improving deployment security.
- Led investigation and implementation efforts involving Model Context Protocol (MCP), JWT-protected services, AgentCore runtime discovery, agent card retrieval, and enterprise authentication patterns.
- Delivered end-to-end solution architecture, cloud infrastructure, backend engineering, DevOps automation, security controls, and production-ready enterprise AI platform capabilities for one of Toptal's diamond-tier enterprise clients.
Technologies: GitLab CI/CD, AWS Cloud Development Kit (CDK), JWT Authentication, Amazon Cognito, Python, AWS Lambda, Serverless Architecture, Model Context Protocol (MCP), AgentCore Registry, AWS Bedrock AgentCore, Bedrock, Solution Architecture, Governance Workflows, DevOps, OpenID Connect (OIDC), Terraform, Cloud Architecture, Enterprise AI
AWS AI Developer
2026 - 2026
Toptal
- Designed and built a production-grade Agentic AI shopping assistant on AWS Bedrock AgentCore, enabling conversational AI, product discovery, personalized recommendations, cart management, and checkout automation for e-commerce workflows.
- Architected a multi-agent system using Strands SDK, Claude Sonnet 4.6, and LLM orchestration, coordinating product research, customer profile management, tool calling, cart operations, and checkout workflows.
- Developed cloud-native AI services with FastAPI, Python, AWS Bedrock, DynamoDB, Amazon S3, and Boto3, delivering scalable backend infrastructure for enterprise AI applications.
- Implemented advanced state management, session persistence, customer memory, and conversation history architecture using DynamoDB, S3 Session Management, and distributed runtime patterns for reliable agent execution.
- Built intelligent knowledge systems combining structured catalog search, Retrieval-Augmented Generation (RAG), Knowledge Bases, and semantic retrieval to improve product recommendations and customer guidance.
- Engineered deterministic tool-calling frameworks for product comparison, cart management, customer profiling, checkout preparation, loyalty programs, and workflow automation.
- Designed structured API contracts and WebSocket-based frontend integrations, enabling dynamic rendering of search, comparison, cart, and checkout experiences across React applications.
- Developed personalized customer experiences using AI-driven recommendation systems, profile management, purchase history tracking, long-term memory, and loyalty-based decision logic.
- Automated deployment and release workflows using Docker, AWS CodeBuild, Amazon S3, containerized applications, and AWS Bedrock AgentCore runtime infrastructure.
- Led end-to-end solution architecture, AI product development, cloud architecture, backend engineering, deployment automation, and production delivery for a large-scale Generative AI commerce platform.
Technologies: Amazon Web Services (AWS), Artificial Intelligence (AI), Machine Learning, Large Language Models (LLMs), Amazon Bedrock, Python, Strands SDK, Claude Sonnet, WebSockets, Boto3, REST APIs, Docker, AWS CodeBuild, Amazon S3 (AWS S3), Amazon DynamoDB, FastAPI
Senior AI Engineer
2023 - 2025
Jazz Pakistan
- Built a prompt-engineered, RAG-based customer-support chatbot for a 70 million+ Telco userbase, cutting live support calls by 30% and delivering $2 million in annual support cost savings.
- Deployed a fully automated email-reply system using n8n, AWS, and Google Workspace to handle thousands of daily inquiries with zero manual touch, eliminating multiple FTEs and saving nearly $300,000 per year.
- Developed a real-time customer pulse dashboard to surface sentiment across social channels, enabling faster customer experience decisions and driving 30% uplift in campaign performance and churn reduction.
- Created AI agents that match talent pools to project job descriptions and integrated voice bots into the sales platform, improving outsourcing efficiency by 25% and unlocking around $500,000 in annual margin and revenue gains.
Technologies: AI Systems, Data Engineering, Data Visualization, Decision Analytics, Business Intelligence (BI), Data Modeling, Docker, ETL Pipelines, AWS (Lambda + API Gateway), Google Workspace, KPI & Performance Analysis, Experimental Design, LangChain, LLM Applications, Machine Learning, PostgreSQL, Predictive Analytics, Python, Python for Analytics, n8n, SQL, Statistical Modeling, Visual Studio Code (VS Code), Workflow Automation, Data Analytics, FastAPI, Git, Large Language Models (LLMs), Artificial Intelligence (AI), Deep Learning, GPS, GPS Tracker, Real-time Systems, Spatial Analysis, Combinatorial Optimization, Computer Vision, Payment APIs, Payments, Data Science, IT Management, Azure, Claude API, Anthropic, Claude, AI Agents, Agentic AI, APIs, Node.js, React, TypeScript, AI Copilots, ChatGPT, ChatGPT API, ChatGPT Prompts, Claude Agent SDK, IT Project Management, IT Projects, Project Management, OpenCV, C++, NVIDIA Jetson, You Only Look Once (YOLO), Edge Computing, Radio Frequency (RF) Protocol, Thermal Imaging, Natural Language Processing (NLP), Prompt Engineering, Android, App Intents, CI/CD Pipelines, Expo, GitHub, Gmail API, Google Calendar, LangGraph, Mobile Applications, OAuth, OpenAI API, Push Notifications, REST APIs, React Native, Speech-to-Text (STT), iOS, Machine Learning Operations (MLOps), Image Processing, Neural Networks, Google Cloud Platform (GCP), AI Programming, AI Tools, Airtable, JavaScript, Vapi, AI Voice Agents, Solution Architecture, Agentic Frameworks, Architecture, Large Language Model Operations (LLMOps), Lovable, RAG Pipelines, Microsoft Copilot, AI Security, Generative Artificial Intelligence (GenAI), Reinforcement Learning from Human Feedback (RLHF), PyTorch, XGBoost, Cybersecurity, Networking, AI-generated Video, Workflows, Data Privacy, Identity & Access Management (IAM), Personally Identifiable Information (PII), Vector Databases, LlamaIndex, Analytics, Snowflake, Tableau, Financial Data, Model Context Protocol (MCP), Pydantic, MongoDB, API Development, Generative Pre-trained Transformers (GPT), OpenAI, AWS Deployment, AWS DevOps, Automation, DevOps, Infrastructure, AI Design, Gemini Enterprise, Vertex AI, JSON, Natural Language Toolkit (NLTK), Data Extraction, Data Pipelines, Data Labeling, Databricks, Gensim, Information Extraction, PySpark, SpaCy, AI Architecture, AI Consulting, AI Integration, Cloud Architecture, Microsoft 365, RAG Architecture, ERP Systems, HubSpot CRM, Odoo, Full-stack Development, Image Generation, Forecasting, System Modeling, Retrieval-augmented Generation (RAG), Agentic RAG Systems, Data Analysis, Attribution Modeling, Customer Segmentation, Funnel Marketing, Markov Model, Marketing, Marketing Attribution, Natural Language Queries, Natural Language Understanding (NLU), A/B Testing, Model Validation, ARIMA, NumPy, Pandas, BigQuery, StatsModels, Multi-agent Systems, System Design, Supabase, Stripe, Data Architecture, Business Rules Engine, Parse, NoSQL, Event-driven Architecture, RAG Systems, Security, Technical Leadership, Data Quality, Transformers, Apache Parquet, Parquet, Scikit-learn, Seaborn, Prefect, Hetzner, Quantitative Analysis, AWS Fargate, AWS IAM, AWS Global Accelerator, AIOps, Claude Code, Cursor AI, Healthcare, Jinja, Optical Character Recognition (OCR), PDF, Azure AI Document Intelligence, Engines, ML Pipelines, Scalability, Semantic Search, Qdrant, FAISS, Weaviate, Front-end Development, Document Processing, Data Governance, Legal, AWS Lambda, Amazon S3 (AWS S3), Cost Reduction & Optimization (Cost-down), Document Parsing, Model Evaluation, Source Code Review, System Architecture, ETL, Jupyter Notebook, Recommendation Systems, Risk Modeling, Risk Models, ETL Tools, Real-time Data, Streaming Data, Website Data Scraping, Full-stack, Custom Models, Pinecone, Finance, Algorithms, CrewAI, Microsoft Copilot Studio, RESTFul APIs, Semantic Kernel (SK), Terraform, Foundry, Low Code, Azure AI Studio, Azure OpenAI Service, C#, GraphQL API, Bicep, Microsoft Power Platform, Azure AI Search, AutoGen, Voice Chat, WhatsApp Business, WhatsApp API, Data Build Tool (dbt), Data Integration, AWS Database Migration Service (DMS), AWS Glue, Cortex, Distributed Systems, Cloud Platforms, Flask, Workflow Automation & System Integration, CRM APIs, Debugging, Platform Design, API Design, AI SDK, AI Model Training, Training, API Integration, WhatsApp, Rule-based Programming, Benchmarking, Hyperparameter Tuning, OpenAI GPT-4 API, Automation Tools, Data Validation, Automations, Calendly API, DocuSign, Email Integration, Zoom, AI Automation, Inventory Management, Invoice Management, Leadership, SDKs, Team Leadership, Databases, Sales Forecasting, Vector Search, Machine Learning (ML) APIs, Chatbot Conversation Design, Chatbots, OpenAI SDK, Platform Engineering, Vercel, Agentic AI Systems, Governance, Regression Testing, Unity Catalog, Decision Modeling, Decision Trees, Enterprise Architecture, Stakeholder Management, Zapier, AI Agent Orchestration, Harness, Light LLMs, Pgvector, MySQL, Data Cleaning, Google Street View, Linear Regression, R, Scaling, Statistical Analysis, Statistical Learning, CRM, Perplexity, Gemini, Whisper, Audio, Coding, Patreon, Fintech, Azure ML Studio, Azure Machine Learning, Markdown, Hooks, AWS Bedrock AgentCore
Data Scientist
2022 - 2023
Bykea
- Pioneered a dynamic pricing engine to incorporate demand-responsive fares, boosting Bykea's profit margins by $4 million.
- Engineered a transformative driver profiling system, significantly augmenting partner income by 30%, enhancing Bykea's unique value proposition.
- Optimized budget allocation and customer targeting using RFM segmentation, increasing engagement by 20% across millions of users, while automating operational processes to reduce data processing time by 30%.
Technologies: SQL, MySQL, Python, AI Systems, Data Analytics, Business Intelligence (BI), Data Visualization, Decision Analytics, Data Engineering, ETL Pipelines, FastAPI, Data Modeling, AWS (Lambda + API Gateway), KPI & Performance Analysis, LangChain, LLM Applications, Machine Learning, Git, Artificial Intelligence (AI), Deep Learning, GPS, GPS Tracker, Real-time Systems, Spatial Analysis, Combinatorial Optimization, Computer Vision, Payment APIs, Payments, IT Management, APIs, Node.js, React, AI Copilots, ChatGPT, ChatGPT API, ChatGPT Prompts, C++, Natural Language Processing (NLP), Prompt Engineering, CI/CD Pipelines, GitHub, Gmail API, Google Calendar, LangGraph, Mobile Applications, Push Notifications, REST APIs, React Native, Speech-to-Text (STT), Image Processing, Neural Networks, Google Cloud Platform (GCP), AI Programming, AI Tools, Airtable, Architecture, RAG Pipelines, Microsoft Copilot, AI Security, Generative Artificial Intelligence (GenAI), Reinforcement Learning from Human Feedback (RLHF), PyTorch, XGBoost, Cybersecurity, Data Privacy, Identity & Access Management (IAM), Personally Identifiable Information (PII), Vector Databases, Analytics, Snowflake, Tableau, Financial Data, Model Context Protocol (MCP), Pydantic, MongoDB, API Development, Automation, DevOps, AI Design, Natural Language Toolkit (NLTK), Data Extraction, Data Pipelines, Data Labeling, Databricks, Gensim, Information Extraction, PySpark, SpaCy, AI Consulting, AI Integration, Microsoft 365, RAG Architecture, HubSpot CRM, Full-stack Development, Image Generation, Forecasting, System Modeling, Agentic RAG Systems, Data Analysis, Attribution Modeling, Customer Segmentation, Funnel Marketing, Markov Model, Marketing, Marketing Attribution, Natural Language Queries, Natural Language Understanding (NLU), A/B Testing, Model Validation, ARIMA, NumPy, Pandas, BigQuery, StatsModels, Multi-agent Systems, System Design, Supabase, Regulatory Compliance, Business Rules Engine, NoSQL, Security, Technical Leadership, Data Quality, Parquet, Seaborn, Prefect, Hetzner, Quantitative Analysis, PDF, Engines, ML Pipelines, Scalability, Semantic Search, Qdrant, FAISS, Document Processing, Data Governance, Cost Reduction & Optimization (Cost-down), Model Evaluation, Source Code Review, System Architecture, Jupyter Notebook, Recommendation Systems, Risk Modeling, Risk Models, Data-informed Recommendations, Full-stack, Custom Models, Finance, Algorithms, RESTFul APIs, Semantic Kernel (SK), Terraform, Foundry, Azure AI Studio, Azure OpenAI Service, C#, GraphQL API, Bicep, Microsoft Power Platform, Azure AI Search, AutoGen, Data Build Tool (dbt), AWS Database Migration Service (DMS), AWS Glue, Workflow Automation & System Integration, Debugging, API Design, Training, Benchmarking, Automation Tools, Data Validation, Automations, Calendly API, DocuSign, Zoom, AI Automation, Logistics, Leadership, Platform Engineering, Agentic AI Systems, Governance, Regression Testing, Unity Catalog, Decision Trees, Zapier, Light LLMs, Data Cleaning, Google Street View, Linear Regression, R, Scaling, Statistical Analysis, Statistical Learning, CRM, Bayesian Statistics, Pytest, Statistics, Streamlit, Datadog, Coding, Google BigQuery, Azure ML Studio, Azure Machine Learning, Markdown, Hooks
Business Analyst
2021 - 2022
Daraz
- Established primary fintech and product analytics at Daraz and managed tracking requirements with Firebase and Google Tag Manager, improving data capture efficiency and driving 15% improvement in nationwide delivery efficiency.
- Led churn analysis to pinpoint UX bottlenecks, delivering targeted fixes that increased customer retention by 3% using Looker/Data Studio.
- Strengthened Daraz’s competitive position by optimizing marketing budgets with RFM profiling, boosting promo and voucher-driven engagement across millions of users.
Technologies: Business Intelligence (BI), Data Engineering, Data Modeling, Data Visualization, Decision Analytics, Data Analytics, ETL Pipelines, AI Systems, Google Workspace, LangChain, LLM Applications, Machine Learning, MySQL, FastAPI, PostgreSQL, Predictive Analytics, KPI & Performance Analysis, Deep Learning, ChatGPT Prompts, GitHub, Gmail API, Google Calendar, Mobile Applications, REST APIs, React Native, Speech-to-Text (STT), RAG Pipelines, PyTorch, XGBoost, Data Privacy, Vector Databases, Analytics, Tableau, Financial Data, Model Context Protocol (MCP), Pydantic, MongoDB, Amazon Web Services (AWS), Automation, DevOps, Data Extraction, Data Labeling, Gensim, Information Extraction, PySpark, Forecasting, Data Analysis, Customer Segmentation, Marketing, Marketing Attribution, StatsModels, NoSQL, Quantitative Analysis, PDF, Engines, Semantic Search, Document Processing, Data Governance, Cost Reduction & Optimization (Cost-down), Model Evaluation, Source Code Review, Jupyter Notebook, Data-informed Recommendations, Data Validation, Governance, Regression Testing, Unity Catalog, Light LLMs, Data Cleaning, Linear Regression, R, Scaling, Statistical Analysis, Statistical Learning, CRM, Bayesian Statistics, Pytest, Statistics, Streamlit, Datadog, Coding, Google BigQuery
Experience
Syngenta AI Foundry AgentCore Registry & AI Governance Platform
PROBLEM
Syngenta, a global agriculture technology leader and one of Toptal's diamond-tier enterprise clients, needed a centralized platform to discover, govern, and manage AI agents and MCP gateways across its AI Foundry ecosystem. Existing AI assets lacked centralized visibility, lifecycle management, approval workflows, and governance controls.
ACTION
Designed and delivered a production-grade AI governance platform using AWS Bedrock AgentCore, AWS Lambda, API Gateway, Amazon Cognito, CloudFront, Amazon S3, Terraform, Serverless Framework, Python, and GitLab CI/CD. Built automated ingestion services for AI agents and MCP gateways, implemented governance workflows, metadata normalization, JWT-based role-based access control, and secure OIDC-based CI/CD pipelines.
RESULT
Delivered a centralized AI registry enabling enterprise-wide discovery, governance, and lifecycle management of AI agents and MCP gateways. Established secure access controls, automated asset discovery, metadata management, and scalable cloud infrastructure for enterprise AI operations. Delivered as part of a strategic production AI initiative for Syngenta, one of Toptal's diamond-tier enterprise clients.
Syngenta, a global agriculture technology leader and one of Toptal's diamond-tier enterprise clients, needed a centralized platform to discover, govern, and manage AI agents and MCP gateways across its AI Foundry ecosystem. Existing AI assets lacked centralized visibility, lifecycle management, approval workflows, and governance controls.
ACTION
Designed and delivered a production-grade AI governance platform using AWS Bedrock AgentCore, AWS Lambda, API Gateway, Amazon Cognito, CloudFront, Amazon S3, Terraform, Serverless Framework, Python, and GitLab CI/CD. Built automated ingestion services for AI agents and MCP gateways, implemented governance workflows, metadata normalization, JWT-based role-based access control, and secure OIDC-based CI/CD pipelines.
RESULT
Delivered a centralized AI registry enabling enterprise-wide discovery, governance, and lifecycle management of AI agents and MCP gateways. Established secure access controls, automated asset discovery, metadata management, and scalable cloud infrastructure for enterprise AI operations. Delivered as part of a strategic production AI initiative for Syngenta, one of Toptal's diamond-tier enterprise clients.
Multi Agent Shopping Intelligence on AWS Bedrock AgentCore
PROBLEM
The client wanted to build a production-grade conversational eCommerce assistant capable of handling intelligent product discovery, contextual recommendations, product comparison, cart workflows, and AI-guided checkout experiences. Existing eCommerce chatbots lacked persistent memory, retrieval grounding, structured orchestration, and reliable transactional handling across multi-step customer journeys.
ACTION
A multi-agent AI commerce platform was designed and deployed on AWS Bedrock AgentCore using Claude Sonnet 4.6, FastAPI, Strands SDK, DynamoDB, S3, and Bedrock Knowledge Base with Titan Embeddings. Orchestration workflows coordinated product search, comparison logic, cart management, checkout execution, retrieval-augmented search, persistent session memory, and structured front-end response handling.
RESULT
The platform enabled an end-to-end conversational shopping experience from product discovery to checkout through a scalable AWS-native architecture. Customers received personalized recommendations, contextual comparison flows, persistent shopping memory, and AI-guided checkout experiences supported by reliable orchestration and production-grade transactional workflows.
The client wanted to build a production-grade conversational eCommerce assistant capable of handling intelligent product discovery, contextual recommendations, product comparison, cart workflows, and AI-guided checkout experiences. Existing eCommerce chatbots lacked persistent memory, retrieval grounding, structured orchestration, and reliable transactional handling across multi-step customer journeys.
ACTION
A multi-agent AI commerce platform was designed and deployed on AWS Bedrock AgentCore using Claude Sonnet 4.6, FastAPI, Strands SDK, DynamoDB, S3, and Bedrock Knowledge Base with Titan Embeddings. Orchestration workflows coordinated product search, comparison logic, cart management, checkout execution, retrieval-augmented search, persistent session memory, and structured front-end response handling.
RESULT
The platform enabled an end-to-end conversational shopping experience from product discovery to checkout through a scalable AWS-native architecture. Customers received personalized recommendations, contextual comparison flows, persistent shopping memory, and AI-guided checkout experiences supported by reliable orchestration and production-grade transactional workflows.
Dynamic Marketplace Optimization & Competitive Intelligence Platform for a Ride-hailing Business
PROBLEM
A ride-hailing business generated millions of daily events across ride requests, bidding, payments, and demand signals, but data was fragmented across systems. Reporting was slow, pricing decisions were reactive, and competitor pricing signals were not systematically incorporated, limiting the ability to optimize marketplace performance.
ACTION
A unified marketplace optimization platform was built to ingest and process high-volume internal operational data alongside external competitor pricing signals. Automated pipelines consolidated millions of daily events into analytics-ready schemas, enabling demand modeling, price sensitivity analysis, and competitive benchmarking through real-time dashboards.
RESULT
Reporting latency dropped significantly, pricing teams gained real-time visibility into demand and competitor movements, and pricing decisions shifted from reactive adjustments to structured, data-driven optimization at scale.
A ride-hailing business generated millions of daily events across ride requests, bidding, payments, and demand signals, but data was fragmented across systems. Reporting was slow, pricing decisions were reactive, and competitor pricing signals were not systematically incorporated, limiting the ability to optimize marketplace performance.
ACTION
A unified marketplace optimization platform was built to ingest and process high-volume internal operational data alongside external competitor pricing signals. Automated pipelines consolidated millions of daily events into analytics-ready schemas, enabling demand modeling, price sensitivity analysis, and competitive benchmarking through real-time dashboards.
RESULT
Reporting latency dropped significantly, pricing teams gained real-time visibility into demand and competitor movements, and pricing decisions shifted from reactive adjustments to structured, data-driven optimization at scale.
Geospatial Socioeconomic Intelligence Platform for Public & Development Programs
PROBLEM
Telecom operators and development organizations lacked reliable, localized socioeconomic insights due to fragmented public, government, and behavioral data sources. Manual research and external consulting made the analysis slow, expensive, and difficult to scale.
ACTION
An end-to-end geospatial intelligence platform was implemented to automate data acquisition, standardize datasets, and enrich signals at neighborhood, district, and city levels. Machine learning models generated predictive wealth indices and segmentation profiles delivered through interactive maps and dashboards.
RESULT
Organizations gained near-instant access to localized demographic and economic intelligence, reducing analysis time from weeks to seconds and enabling more precise targeting, planning, and resource allocation at a national scale.
Telecom operators and development organizations lacked reliable, localized socioeconomic insights due to fragmented public, government, and behavioral data sources. Manual research and external consulting made the analysis slow, expensive, and difficult to scale.
ACTION
An end-to-end geospatial intelligence platform was implemented to automate data acquisition, standardize datasets, and enrich signals at neighborhood, district, and city levels. Machine learning models generated predictive wealth indices and segmentation profiles delivered through interactive maps and dashboards.
RESULT
Organizations gained near-instant access to localized demographic and economic intelligence, reducing analysis time from weeks to seconds and enabling more precise targeting, planning, and resource allocation at a national scale.
Real-time Commercial Performance & Transaction Intelligence Platform for a Global eCommerce
PROBLEM
A global eCommerce marketplace processed millions of transactions daily, but commercial data was fragmented across systems, leading to delayed revenue insights, spreadsheet-driven reporting, and limited real-time visibility for leadership.
ACTION
A cloud-native analytics platform was designed to continuously ingest transactional data, standardize datasets via scalable ETL pipelines, and compute consistent revenue and sales KPIs across markets. Interactive dashboards provided real-time visibility for executives and operations.
RESULT
Reporting cycles shifted from hours to real-time, leadership gained faster insight into performance changes, manual reporting overhead declined, and the platform scaled with growing transaction volumes and markets.
A global eCommerce marketplace processed millions of transactions daily, but commercial data was fragmented across systems, leading to delayed revenue insights, spreadsheet-driven reporting, and limited real-time visibility for leadership.
ACTION
A cloud-native analytics platform was designed to continuously ingest transactional data, standardize datasets via scalable ETL pipelines, and compute consistent revenue and sales KPIs across markets. Interactive dashboards provided real-time visibility for executives and operations.
RESULT
Reporting cycles shifted from hours to real-time, leadership gained faster insight into performance changes, manual reporting overhead declined, and the platform scaled with growing transaction volumes and markets.
4G Customer Churn Prediction Model
PROBLEM
As 4G adoption increased, customer churn became harder to manage. Retention efforts were reactive, triggered too late, and relied on broad incentives that led to inefficient spending and missed intervention opportunities.
ACTION
A machine-learning churn prediction model was developed using customer behavior, usage patterns, and engagement signals. Individual-level churn risk scores were integrated directly into CRM and retention workflows for proactive intervention.
RESULT
Retention teams acted earlier on high-risk customers, reduced unnecessary incentives for stable users, and improved overall retention effectiveness without increasing operational complexity.
As 4G adoption increased, customer churn became harder to manage. Retention efforts were reactive, triggered too late, and relied on broad incentives that led to inefficient spending and missed intervention opportunities.
ACTION
A machine-learning churn prediction model was developed using customer behavior, usage patterns, and engagement signals. Individual-level churn risk scores were integrated directly into CRM and retention workflows for proactive intervention.
RESULT
Retention teams acted earlier on high-risk customers, reduced unnecessary incentives for stable users, and improved overall retention effectiveness without increasing operational complexity.
LLM-powered Customer Support Chatbot
PROBLEM
Customer support teams handled extremely high volumes of repetitive queries, consuming live agent capacity and slowing response times for complex issues. Existing knowledge bases were not accessible through conversational interfaces.
ACTION
An LLM-powered chatbot using a retrieval-augmented generation architecture was deployed to combine approved knowledge sources with structured customer data. Confidence-based escalation routed uncertain queries to human agents.
RESULT
A large share of inbound volume was automated, response times improved, support costs declined, and agents focused on high-value cases without compromising service quality.
Customer support teams handled extremely high volumes of repetitive queries, consuming live agent capacity and slowing response times for complex issues. Existing knowledge bases were not accessible through conversational interfaces.
ACTION
An LLM-powered chatbot using a retrieval-augmented generation architecture was deployed to combine approved knowledge sources with structured customer data. Confidence-based escalation routed uncertain queries to human agents.
RESULT
A large share of inbound volume was automated, response times improved, support costs declined, and agents focused on high-value cases without compromising service quality.
AI Agents for Talent Matching
PROBLEM
A consulting platform relied on manual, inconsistent talent-to-project matching across large, global talent pools, resulting in slow placements, missed opportunities, and high staffing overhead.
ACTION
AI agents were developed to analyze project requirements and consultant profiles, generating scored and ranked recommendations with explainable rationale while keeping humans in the loop for final decisions.
RESULT
Time-to-match decreased, placement consistency improved, talent utilization increased, and staffing teams handled higher volumes without additional headcount.
A consulting platform relied on manual, inconsistent talent-to-project matching across large, global talent pools, resulting in slow placements, missed opportunities, and high staffing overhead.
ACTION
AI agents were developed to analyze project requirements and consultant profiles, generating scored and ranked recommendations with explainable rationale while keeping humans in the loop for final decisions.
RESULT
Time-to-match decreased, placement consistency improved, talent utilization increased, and staffing teams handled higher volumes without additional headcount.
AI Voice Agents for Sales Enablement
PROBLEM
Sales teams spent excessive time on lead qualification, follow-ups, and routing, leading to slow responses, missed leads, and inefficient use of senior sales capacity.
ACTION
AI voice agents were deployed to autonomously conduct qualification calls, capture intent and urgency, execute follow-ups, and route qualified leads into existing sales workflows.
RESULT
Response speed improved, lead engagement became consistent, conversion efficiency increased, and sales teams focused on high-intent opportunities without expanding headcount.
Sales teams spent excessive time on lead qualification, follow-ups, and routing, leading to slow responses, missed leads, and inefficient use of senior sales capacity.
ACTION
AI voice agents were deployed to autonomously conduct qualification calls, capture intent and urgency, execute follow-ups, and route qualified leads into existing sales workflows.
RESULT
Response speed improved, lead engagement became consistent, conversion efficiency increased, and sales teams focused on high-intent opportunities without expanding headcount.
Customer Virtual Assistant with Voice-enabled Humanoid Interface
PROBLEM
Digital banking websites struggled to build trust and explain products through static pages and text-based chat, pushing early-stage users toward call centers.
ACTION
A voice-enabled, 3D humanoid virtual assistant was embedded into the website, enabling natural voice interaction and guided product discovery with a human-like front-desk experience.
RESULT
User engagement and product understanding improved, discovery friction decreased, and the organization delivered a scalable, trust-building digital experience without increasing reliance on live support.
Digital banking websites struggled to build trust and explain products through static pages and text-based chat, pushing early-stage users toward call centers.
ACTION
A voice-enabled, 3D humanoid virtual assistant was embedded into the website, enabling natural voice interaction and guided product discovery with a human-like front-desk experience.
RESULT
User engagement and product understanding improved, discovery friction decreased, and the organization delivered a scalable, trust-building digital experience without increasing reliance on live support.
Global Gender Analytics & Inference Deployment
PROBLEM
Gender-disaggregated data was incomplete or unavailable across markets, limiting inclusive product design and responsible targeting. Existing approaches were fragmented and difficult to scale.
ACTION
A containerized, privacy-preserving gender inference model was deployed across large-scale telecom data pipelines using probabilistic signals rather than explicit identifiers.
RESULT
Organizations gained consistent, ethical gender analytics at a national scale, enabling inclusion initiatives, improved engagement, and measurable commercial impact.
Gender-disaggregated data was incomplete or unavailable across markets, limiting inclusive product design and responsible targeting. Existing approaches were fragmented and difficult to scale.
ACTION
A containerized, privacy-preserving gender inference model was deployed across large-scale telecom data pipelines using probabilistic signals rather than explicit identifiers.
RESULT
Organizations gained consistent, ethical gender analytics at a national scale, enabling inclusion initiatives, improved engagement, and measurable commercial impact.
RFM-based Voucher & Marketing Optimization
PROBLEM
Voucher spending relied on broad rules, causing discount leakage to loyal users and inefficient incentives for promo-driven segments with unclear incremental impact.
ACTION
An RFM-based segmentation framework grouped users into interpretable cohorts, each mapped to differentiated voucher strategies aligned with expected behavioral outcomes.
RESULT
Marketing ROI improved, unnecessary discounts declined, and incentive spend shifted toward cohorts where promotions drove measurable incremental behavior.
Voucher spending relied on broad rules, causing discount leakage to loyal users and inefficient incentives for promo-driven segments with unclear incremental impact.
ACTION
An RFM-based segmentation framework grouped users into interpretable cohorts, each mapped to differentiated voucher strategies aligned with expected behavioral outcomes.
RESULT
Marketing ROI improved, unnecessary discounts declined, and incentive spend shifted toward cohorts where promotions drove measurable incremental behavior.
Location-based Fraud Detection System for Mobility Operations
PROBLEM
As urban scale increased, GPS spoofing, fake rides, and incentive abuse caused financial losses. Rule-based controls detected fraud only after payouts were issued.
ACTION
A real-time fraud detection pipeline analyzed live GPS and trip data using spatial feature engineering and machine-learning anomaly detection, surfacing risk before payouts.
RESULT
Fraud losses declined, response times improved, and fraud control shifted from post-hoc recovery to proactive prevention.
As urban scale increased, GPS spoofing, fake rides, and incentive abuse caused financial losses. Rule-based controls detected fraud only after payouts were issued.
ACTION
A real-time fraud detection pipeline analyzed live GPS and trip data using spatial feature engineering and machine-learning anomaly detection, surfacing risk before payouts.
RESULT
Fraud losses declined, response times improved, and fraud control shifted from post-hoc recovery to proactive prevention.
AI-first Operations Transformation
PROBLEM
Operations relied on manual coordination across spreadsheets, documents, and messaging tools, creating delays, errors, and scalability limits without hiring more staff.
ACTION
Core workflows were redesigned around an AI-first execution layer using agents and automation to orchestrate work end-to-end with event-driven coordination.
RESULT
Operational friction declined, throughput scaled without headcount growth, and teams achieved sustained productivity and reliability gains.
Operations relied on manual coordination across spreadsheets, documents, and messaging tools, creating delays, errors, and scalability limits without hiring more staff.
ACTION
Core workflows were redesigned around an AI-first execution layer using agents and automation to orchestrate work end-to-end with event-driven coordination.
RESULT
Operational friction declined, throughput scaled without headcount growth, and teams achieved sustained productivity and reliability gains.
AI-powered Hiring & Candidate Assessment System
PROBLEM
Hiring workflows relied on manual screening, fragmented assessments, and recruiter-heavy coordination, leading to slow cycles and inconsistent evaluations as volume grew.
ACTION
An end-to-end AI-driven hiring system automated screening, assessments, interviews, decisions, and onboarding using LLM-based evaluation integrated with ATS workflows.
RESULT
Time-to-hire shortened, consistency improved, candidate experience strengthened, and hiring scaled without increasing recruiter headcount.
Hiring workflows relied on manual screening, fragmented assessments, and recruiter-heavy coordination, leading to slow cycles and inconsistent evaluations as volume grew.
ACTION
An end-to-end AI-driven hiring system automated screening, assessments, interviews, decisions, and onboarding using LLM-based evaluation integrated with ATS workflows.
RESULT
Time-to-hire shortened, consistency improved, candidate experience strengthened, and hiring scaled without increasing recruiter headcount.
Customer Service Performance & Cost Optimization
PROBLEM
High ticket volumes and limited agent capacity drove rising costs, slow resolution times, and overloaded queues, while skilled agents were spread across low- and high-impact issues.
ACTION
Decision-grade analytics using Pareto analysis, productivity diagnostics, and ticket classification redesigned routing and prioritization policies.
RESULT
Support costs declined, resolution times improved, agent utilization increased, and high-impact cases received focused human attention.
High ticket volumes and limited agent capacity drove rising costs, slow resolution times, and overloaded queues, while skilled agents were spread across low- and high-impact issues.
ACTION
Decision-grade analytics using Pareto analysis, productivity diagnostics, and ticket classification redesigned routing and prioritization policies.
RESULT
Support costs declined, resolution times improved, agent utilization increased, and high-impact cases received focused human attention.
Education
2019 - 2020
Master’s Degree in Business Analytics
Karachi School of Business and Leadership (KSBL) - Karachi, Pakistan
2016 - 2019
Bachelor's Degree in Economics
International Islamic University Islamabad (IIUI) - Islamabad, Pakistan
Skills
Libraries/APIs
Node.js, React, Gmail API, OpenAI API, REST APIs, PyTorch, XGBoost, Pydantic, API Development, Natural Language Toolkit (NLTK), PySpark, SpaCy, NumPy, Pandas, Stripe, Scikit-learn, Salesforce REST API, Salesforce API, Salesforce Bulk API, GraphQL API, WhatsApp API, Calendly API, Google Street View, Dlib, OpenAI Assistants API, Claude API, OpenCV
Tools
Git, n8n, Claude, ChatGPT, Claude Agent SDK, GitHub, Microsoft Copilot, Tableau, AWS Deployment, Gensim, Odoo, ARIMA, BigQuery, StatsModels, Seaborn, Prefect, AWS Fargate, AWS IAM, Salesforce DX, Claude Code, Terraform, Azure OpenAI Service, AWS Glue, AI SDK, Zoom, Amazon QuickSight, Zapier, Whisper, Pytest, Azure ML Studio, Azure Machine Learning, NVIDIA Jetson, You Only Look Once (YOLO), Expo, Google Workspace, AWS CodeBuild, GitLab CI/CD, AWS Cloud Development Kit (CDK), Amazon Cognito, Amazon CloudFront
Languages
Python, SQL, TypeScript, C++, JavaScript, Snowflake, C#, Bicep, R, Markdown
Frameworks
LangGraph, Agentic Frameworks, LlamaIndex, Business Rules Engine, Jinja, AutoGen, Flask, Streamlit, App Intents, React Native, Metal, Bedrock
Paradigms
Real-time Systems, Model Context Protocol (MCP), Automation, DevOps, ETL, Rule-based Programming, Asynchronous Programming, Radio Frequency (RF) Protocol, Event-driven Architecture, HIPAA Compliance, Business Intelligence (BI), UX Design, Anomaly Detection, Serverless Architecture
Platforms
Docker, Azure, Google Cloud Platform (GCP), Amazon Web Services (AWS), Vertex AI, Databricks, Parse, CrewAI, AWS Lambda, Jupyter Notebook, Kubernetes, Microsoft Copilot Studio, Azure AI Studio, Microsoft Power Platform, Azure AI Search, Cortex, Vercel, Harness, Patreon, Android, iOS, Visual Studio Code (VS Code)
Storage
PostgreSQL, MySQL, Data Integration, MongoDB, JSON, Data Pipelines, NoSQL, Apache Parquet, Amazon S3 (AWS S3), Data Validation, Databases, Datadog, Amazon DynamoDB
Industry Expertise
Project Management, Cybersecurity, Marketing, Healthcare
Other
LangChain, FastAPI, Data Analytics, Statistical Modeling, Predictive Analytics, Machine Learning, Data Engineering, ETL Pipelines, Data Modeling, Workflow Automation, Data Science, Dashboards, Web Scraping, API Integration, Large Language Models (LLMs), Retrieval-augmented Generation (RAG), Natural Language Processing (NLP), Prompt Engineering, AI Voice Agents, Workflow Automation & System Integration, Data Analysis, Artificial Intelligence (AI), Deep Learning, GPS, GPS Tracker, Computer Vision, IT Management, AI Agents, Agentic AI, APIs, AI Copilots, ChatGPT API, ChatGPT Prompts, IT Project Management, IT Projects, Google Calendar, Mobile Applications, Push Notifications, Speech-to-Text (STT), Machine Learning Operations (MLOps), Image Processing, Neural Networks, AI Programming, AI Tools, Airtable, Architecture, Large Language Model Operations (LLMOps), Lovable, RAG Pipelines, AI Security, Generative Artificial Intelligence (GenAI), Reinforcement Learning from Human Feedback (RLHF), Networking, AI-generated Video, Workflows, Data Privacy, Identity & Access Management (IAM), Personally Identifiable Information (PII), Vector Databases, Analytics, Generative Pre-trained Transformers (GPT), OpenAI, AWS DevOps, Infrastructure, AI Design, Gemini Enterprise, Data Extraction, Data Labeling, Information Extraction, AI Architecture, AI Consulting, AI Integration, Cloud Architecture, Microsoft 365, RAG Architecture, ERP Systems, HubSpot CRM, Full-stack Development, Image Generation, Forecasting, System Modeling, Agentic RAG Systems, Attribution Modeling, Customer Segmentation, Funnel Marketing, Markov Model, Marketing Attribution, Natural Language Queries, Natural Language Understanding (NLU), A/B Testing, Model Validation, Multi-agent Systems, System Design, Supabase, Regulatory Compliance, Data Architecture, RAG Systems, Technical Leadership, Data Quality, Transformers, Parquet, Hetzner, Quantitative Analysis, AWS Global Accelerator, AIOps, Salesforce Apex, Salesforce CTI, Salesforce Connect, Cursor AI, Optical Character Recognition (OCR), PDF, Engines, ML Pipelines, Scalability, Semantic Search, Qdrant, FAISS, Weaviate, Front-end Development, Document Processing, Data Governance, Legal, Cost Reduction & Optimization (Cost-down), Document Parsing, Model Evaluation, Source Code Review, System Architecture, AI Chatbots, Amazon Bedrock, Recommendation Systems, Risk Modeling, Risk Models, Data-informed Recommendations, ETL Tools, Real-time Data, Streaming Data, Website Data Scraping, Full-stack, Custom Models, Pinecone, Finance, Algorithms, RESTFul APIs, Semantic Kernel (SK), Foundry, Low Code, Voice Chat, WhatsApp Business, Data Build Tool (dbt), AWS Database Migration Service (DMS), Financial Analysis, Reporting, Distributed Systems, Back-end, Cloud Platforms, CRM APIs, Debugging, Platform Design, API Design, AI Model Training, Training, MLflow, Scraping, Real Estate, WhatsApp, Benchmarking, Hyperparameter Tuning, OpenAI GPT-4 API, Automation Tools, Automations, DocuSign, Email Integration, Webhooks, AI Automation, Consulting, Logistics, Dagster, Inventory Management, Invoice Management, Leadership, SDKs, Team Leadership, Sales Forecasting, Vector Search, Machine Learning (ML) APIs, Chatbot Conversation Design, Chatbots, OpenAI SDK, Platform Engineering, Agentic AI Systems, Governance, Regression Testing, Unity Catalog, Decision Modeling, Decision Trees, Enterprise Architecture, Stakeholder Management, AI Agent Orchestration, Light LLMs, Pgvector, Data Cleaning, Linear Regression, Scaling, Statistical Analysis, Statistical Learning, CRM, Perplexity, Gemini, Audio, Facial Recognition, Facial Tracking, CCTV, Video Surveillance, DeepFace, AdaFace, FaceNet, InsightFace, MagFace, Bayesian Statistics, Statistics, Coding, Google BigQuery, Fintech, Hooks, AWS Bedrock AgentCore, Spatial Analysis, Combinatorial Optimization, Payment APIs, Payments, Anthropic, Edge Computing, Thermal Imaging, CI/CD Pipelines, OAuth, Vapi, Solution Architecture, Financial Data, Security, Azure AI Document Intelligence, ArcFace, CosFace, FaceNet512, GhostFaceNets, AWS (Lambda + API Gateway), Data Visualization, Python for Analytics, Experimental Design, Decision Analytics, KPI & Performance Analysis, AI Systems, LLM Applications, Pricing Optimization, Market Place Analytics, Competitive Strategy, Demand Forecasting, Geospatial Analytics, Predictive Modeling, Feature Engineering, Real-time Analytics, Transaction Analytics, Revenue Analytics, Cloud Data Platforms, Visualization, Churn Modeling, Customer Lifecycle Modeling, Classification, Model Deployment, Conversational AI, AI System Design, Automation & Workflow Integration, Talent Matching Systems, Ranking & Scoring Models, Feature Extraction, Sales Process Automation, Speech Recognition & Synthesis, Multimodal AI Systems, Machine Learning Inference, Privacy-Preserving Analytics, Large-Scale Data Pipelines, Linux Infrastructure, Responsible AI, RFM Segmentation, User Behavior, Marketing Analytics, SQL-Based Data Pipelines, Cohort Analysis, Fraud Detection, Real-time Processing, Operational Dashboards, Process Automation, Operations Design, Low-Code / No-Code Platforms, System Integration, Operational Analytics, Hiring Automation, Decision Support Systems (DSS), Applicant Tracking Systems (ATS), Customer Support Analytics, Cost Optimization, Ticket Classification, Agent Productivity Analysis, Strands SDK, Claude Sonnet, WebSockets, Boto3, JWT Authentication, AgentCore Registry, Governance Workflows, OpenID Connect (OIDC), Enterprise AI, API Gateways
Collaboration That Works
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
1
Share your needs
Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2
Choose your talent
Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3
Start your risk-free talent trial
Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.
Top talent is in high demand.
Start hiring