Muaz is available for hire
Hire MuazMuaz Maqbool
Verified Expert in Engineering
Artificial Intelligence Developer
Lahore, Punjab, Pakistan
Toptal member since June 8, 2023
Bio
Muaz has 8+ years of product development experience, having co-founded multiple AI startups that have raised over $1 million in funding. As a former VP of AI and C-suite consultant, he's delivered 50+ GenAI POCs for clients, including J&J, SBI Growth Advisory, and automotive AI platforms. With 11 published papers and an M.S. in Machine Learning from Georgia Tech, Muaz maintains 30+ engagements with US enterprises spanning LLM architectures, multi-agent systems, RAG, and cloud deployments.
Portfolio
Self-employed
Artificial Intelligence (AI), Computer Vision, Machine Learning...
Spencer Scalzitti
Real Estate, Python, Artificial Intelligence (AI)...
Sales Benchmark Index, LLC
Python, React, Large Language Models (LLMs)...
Experience
- PyTorch - 6 years
- Data Science - 6 years
- Computer Vision - 6 years
- Machine Learning - 6 years
- Python - 6 years
- Deep Learning - 6 years
- Natural Language Processing (NLP) - 5 years
- Generative Pre-trained Transformer 3 (GPT-3) - 1 year
Preferred Environment
Artificial Intelligence (AI), Visual Studio Code (VS Code), Computer Vision, Natural Language Processing (NLP), Data Science, Agentic AI, AI Agents, Generative Artificial Intelligence (GenAI), Technical Leadership
The most amazing...
...thing I’ve developed is a production AI platform that automates how sales consultants analyze enterprise data, from hypothesis to report, fully autonomously.
Work Experience
Computer Vision and NLP Expert | AI Consultant and Contractor
2022 - PRESENT
Self-employed
- Served as a consultant and contractor. Brought in technical expertise and managerial skills to assist 20+ US companies in seamlessly integrating AI into their products.
- Built and maintained over 25 long-term engagements as an expert in the Expert Vetted Batch (Top 1%) category with a remarkable 99% job success rate.
- Leveraged OpenAI LLMs to deliver solutions, including automated property description generation, chatbot development, and embedding-based engines. Created models for medical insurance billing, contributing to the success of multiple businesses.
Technologies: Artificial Intelligence (AI), Computer Vision, Machine Learning, Natural Language Processing (NLP), Data Science, GitHub, Machine Learning Operations (MLOps), AWS Command Line Interface (CLI), Google Cloud Platform (GCP), Language Models, FastAPI, Flask, Generative Pre-trained Transformer 3 (GPT-3), OpenAI GPT-4 API, OpenAI GPT-3 API, OpenCV, PyTorch, ChatGPT, Generative Pre-trained Transformers (GPT), Pandas, Jupyter, Programming, Integration, Natural Language Toolkit (NLTK), SpaCy, LangChain, Amazon Web Services (AWS), Image Recognition, Visualization, Image Processing, Web Scraping, Back-end Development, Google Vision API, Image Search, Labeling, Amazon Rekognition, Fashion, PostgreSQL, Recurrent Neural Networks (RNNs), Neural Networks, Text to Image, Convolutional Neural Networks (CNNs), You Only Look Once (YOLO), Open Neural Network Exchange (ONNX), Regression Modeling, Quantitative Analysis, Forecasting, FFmpeg, Research, Communication, Generative Adversarial Networks (GANs), OpenAI, Custom Models, Generative Artificial Intelligence (GenAI), Architecture, Industrial Internet of Things (IIoT), Time Series Analysis, Transformer Models, Time Series, Retrieval-augmented Generation (RAG), Signal Processing, Data Extraction, PDF, Databases, Cloud, Software Architecture, Algorithms, Models, Data, AI Integration, Prompt Engineering, AI Agents, Rapid Prototyping, Predictive Analytics, System Design, Generative Pre-trained Transformer 4 (GPT-4), Large Language Model Operations (LLMOps), Minimum Viable Product (MVP), Firebase, Data Engineering, OpenAI API, Sentiment Analysis, Azure, AI Design, Optical Character Recognition (OCR), AI Consulting, Geospatial Data, icr, Machine Learning Algorithms, Amazon Machine Learning, Fine-tuning, Amazon SageMaker, AI Programming, Visual Studio Code (VS Code), Docker, Text to Image AI, Supervised Learning, Unsupervised Learning, Hugging Face, API Integration, Benchmarking, Hyperparameter Tuning, Rule-based Programming, Vertex AI, Data Management, Data Governance, Data Pipelines, Snowflake, Claude Code, Startups, Back-end, Vercel, Decision Modeling, LangGraph, Agentic RAG Systems, Decision Trees, Clustering Algorithms, Predictive Modeling, Prediction Markets, AI Tools, Twilio, VoIP, Computer Vision Algorithms, Residual Neural Networks (ResNets), Hugging Face Transformers, NVIDIA TensorRT, DevOps, Real-time Systems, RAG Pipelines, Whisper, Audio, Facial Recognition, Facial Tracking, CCTV, Video Surveillance, Dlib, Edge Computing, Linear Regression, Bayesian Statistics, Pytest, Streamlit, NumPy, Data Labeling, AI Model Training, YOLOv5, Object Tracking, Motion Tracking, Object Detection, Kalman Filtering, Machine Learning (ML) APIs, Azure Machine Learning, Azure ML Studio, Workflow Automation, NX CAD, Agentic AI Systems, RAG Architecture, Best Practices, Bots, n8n, Workflow Automation & System Integration, Redis, Stripe, AI Product Strategy, Synthetic Data Generation, Scripting, AI Engineering, Model Context Protocol (MCP), Business Intelligence (BI), Video Analytics, Multi-agent Systems, Event-driven Architecture, Financial System Implementation, Solution Architecture, Azure Functions, Azure Function App, Amazon Bedrock AgentCore, Knowledge Graphs, LLM Reasoning, AI Architecture, Context Engineering, Fintech, Medical Imaging, 3D Reconstruction, DICOM, Dental Care, Graph Neural Networks (GNNs), Pattern Recognition, Time Series Forecasting, 3D Pose Estimation, Motion Capture, Pose Estimation, HIPAA, Health, Healthcare, Neuroscience, BigQuery, Delta Lake, Google Cloud Storage, LSTM, Feature Engineering, Model Deployment, Model Development, Monitoring, Random Forests, XGBoost, IT Strategy, Enterprise Architecture, Enterprise SaaS, Data Analysis, Agentic Workflow Design, Azure OpenAI Service, Strategy
AI/Finance/Real Estate Expert
2025 - 2026
Spencer Scalzitti
- Built a GenAI-powered outreach cadence engine that automatically generated and sent personalized email and SMS sequences to property leads, dynamically adjusting follow-up chains based on recipient responses and engagement signals.
- Designed and implemented a tiered lead bucketing system that classified incoming responses and routed leads into appropriate cadence tracks, enabling automated re-engagement without manual intervention.
- Developed AI-generated email templates using LLMs that personalized messaging based on property details, ownership data, and buyer tier—replacing manual copywriting across campaigns.
- Built a property ownership data scraping pipeline that aggregated owner records from multiple third-party sources, including Endato, enriching the CRM database with verified contact and ownership information.
- Deployed the full back end as a Cloud Run job on GCP with a PostgreSQL database, migrating from the previous infrastructure to a production FastAPI-based architecture.
- Created a CRM-style property management platform that tracked ownership records and enriched property data through third-party API integrations.
Technologies: Real Estate, Python, Artificial Intelligence (AI), Generative Artificial Intelligence (GenAI), OpenAI, Scraping, Web Scraping, Data Architecture, Risk Assessment, Data Analytics, Underwriting, Credit Underwriting, Finance, Data Science, Data Pipelines, Claude Code, Startups, Back-end, Vercel, Decision Modeling, LangGraph, Agentic RAG Systems, Prediction Markets, AI Tools, AI Automation, AI Agent Orchestration, Graphics Processing Unit (GPU), Hugging Face Transformers, DevOps, RAG Pipelines, Gemini, NumPy, Machine Learning (ML) APIs, RAG Architecture, Vector Search, AI Assistants, Best Practices, AI Product Strategy, AI Engineering, Financial System Implementation, Solution Architecture, LLM Reasoning, AI Architecture, Fintech, Google Cloud Storage, Model Development, Monitoring, Enterprise Architecture, Data Analysis, Agentic Frameworks, Agentic Workflow Design, Strategy
Senior AI Engineer – LLMs, RAG, and Agentic Architectures
2025 - 2026
Sales Benchmark Index, LLC
- Built a full AI workflow platform from scratch supporting three agent architectures (ReAct, Plan-and-Execute, and Fixed Flows), with interruptible human-in-the-loop controls and scheduled execution.
- Migrated the entire AI back end from a legacy stack to a self-hosted FastAPI-based runtime with MCP tool servers, supporting multi-model routing across OpenAI, Anthropic, and Gemini APIs.
- Designed and implemented the SAHF (SBI Analytical Hypothesis Framework), an adaptive looping agent that auto-generates, tests, and refines analytical hypotheses against SBI's proprietary methodology.
- Engineered real-time streaming over SSE and WebSockets across three transport modes, enabling drag-and-drop workflow automation via dynamic graph specifications.
- Built a SkillRouter layer that dynamically routes user requests to specialized AI skills, reducing prompt engineering overhead and improving response accuracy across 50+ GenAI proofs-of-concept.
Technologies: Python, React, Large Language Models (LLMs), Retrieval-augmented Generation (RAG), AI Agents, GraphRAG, LangChain, LangGraph, REST APIs, Data Pipelines, Claude Code, Back-end, Vercel, Decision Modeling, Agentic RAG Systems, Decision Trees, Predictive Modeling, AI Tools, AI Agent Orchestration, Light LLMs, Harness, Pgvector, DevOps, vLLM, LLM Integration, RAG Pipelines, Gemini, Linear Regression, Streamlit, Technical Leadership, Machine Learning (ML) APIs, Workflow Automation, Agentic AI Systems, RAG Architecture, Vector Search, AI Assistants, Best Practices, Bots, n8n, Workflow Automation & System Integration, Redis, AI Engineering, Model Context Protocol (MCP), Business Intelligence (BI), Multi-agent Systems, Solution Architecture, LLM Reasoning, AI Architecture, Context Engineering, Google Cloud Storage, Model Development, Model Evaluation, Enterprise Architecture, Data Analysis, Agentic Frameworks, Agentic Workflow Design
Vice President of Artificial Intelligence
2024 - 2026
IDPrivacy
- Led AI architecture and delivery for a privacy-focused agentic communication platform scaled to 100+ automotive dealerships. Managed a 10-person AI team across the full product lifecycle, from technical hiring through production deployment.
- Architected LangChain ReAct agents and RAG pipelines for scalable Text-to-SQL interactions across dealership CRM and inventory data.
- Built multi-modal agentic workflows for voice, SMS, and email using n8n, integrated with 11Labs and Retell for conversational voice agents.
- Developed API wrappers for the top 5 U.S. automotive CRMs, integrating lead, inventory, and service data into AI agents for full-cycle customer journey automation.
- Achieved certified OEM integration status for Nissan, Mitsubishi, and Hyundai across 100+ dealership deployments.
- Built a post-call analytics platform with Redis Streams-based RPA workers on Kubernetes and a three-layer LLM prompt architecture for structured call evaluation.
- Fine-tuned Qwen 2.5 with QLoRA on automotive transaction data. Set up CI/CD for model deployments and KPI dashboards for executive reporting.
Technologies: LangChain, LangGraph, Python 3, FastAPI, Redis, Kubernetes, Docker, n8n, OpenAI, PostgreSQL, RAG Systems, ReAct Agents, Qwen, LoRa, ElevenLabs Solutions, Retell AI, ISO 27001, SOC 2, Government Contracting, Agentic Frameworks, Agentic Workflow Design, Azure OpenAI Service, Strategy
LLM AI Expert
2025 - 2025
HMFT Inc.
- Created a scraping web app to scrape 35,000 documents from the public HMFT website.
- Compared different vision-language model solutions to find the best structured data extraction from complex assembly manuals, including Mistral OCR, Google Vision API, AWS Recognition, and EasyOCR.
- Created and deployed a production-grade RAG pipeline using Google Vertex AI Vector Store.
- Deployed the solution in a scalable manner using the GCP Cloud Run service.
Technologies: Artificial Intelligence (AI), Python, Large Language Models (LLMs), AI Chatbots, OpenAI, Chatbots, Data Scraping, Web Scraping, Automation, Data Pipelines, Natural Language Processing (NLP), Supervised Learning, Unsupervised Learning, Hugging Face, RAG Systems, Hyperparameter Tuning, Vector Databases, Startups, Back-end, Vercel, LangGraph, Website Data Scraping, AI Agent Orchestration, Light LLMs, Harness, Pgvector, Hugging Face Transformers, DevOps, LLM Integration, Agentic AI Systems, Vector Search, Best Practices, Bots, Scripting, AI Engineering, LLM Reasoning, Context Engineering, Model Evaluation
AI Developer
2025 - 2025
Higher Level Legal LLC
- Explored the CLIO legal CRM for external app integrations.
- Wrote API wrappers to integrate the CLIO CRM app into a GenAI pipeline.
- Researched legal document automation use cases with regard to AI.
Technologies: Artificial Intelligence (AI), Python, Large Language Models (LLMs), OpenAI, OpenAI GPT-4 API, API Integration, Zapier, Supervised Learning, Vector Databases, Startups, Decision Modeling, Hugging Face Transformers, Streamlit, AI Engineering
LLM Expert
2024 - 2025
Gridwise, Inc
- Built an LLM-powered chat agent that queried a 100,000+ user database, generated dynamic SQL, and delivered personalized gig-economy analytics with sub-500 ms response times.
- Conducted a comparative evaluation of open-source LLMs (Llama 2 and Mistral) versus managed LLMs (Vertex AI PaLM and OpenAI GPT), selecting models that improved inference accuracy by 15%.
- Created a RAG pipeline on Vertex AI for dynamic few-shot test-to-SQL, achieving a 60% accuracy improvement.
- Benchmarked proprietary LLMs (GPT-4o vs. Gemini 2.0 Flash) to select the best model for inference.
- Optimized embeddings to minimize costs without sacrificing performance.
Technologies: Artificial Intelligence (AI), Large Language Models (LLMs), Retrieval-augmented Generation (RAG), Rapid Prototyping, AI Agents, LangChain, Machine Learning Operations (MLOps), Reinforcement Learning from Human Feedback (RLHF), Multi-agent Systems, Data Analytics, Full-stack Development, Agentic AI, Open-source LLMs, RAG Systems, API Integration, Benchmarking, Rule-based Programming, Vector Databases, Snowflake, AI Agent Orchestration, Light LLMs, Hugging Face Transformers, DevOps, LLM Integration, Pytest, AI Assistants, LLM Reasoning, Context Engineering
Python Developer | AI & LangChain Expert
2024 - 2024
Marwan Ezzat
- Developed and deployed a comprehensive pipeline using LangChain and various scraping libraries to extract, summarize, and structure legal rulings from European Parliament websites.
- Automated the generation of daily email summaries in a fixed format, enhancing efficiency and ensuring consistency in legal ruling notifications for stakeholders.
- Deployed the entire system successfully on DigitalOcean, streamlining the process from data extraction to email dissemination, which provided a reliable and scalable solution for legal updates.
Technologies: Python, JavaScript, LangChain, Integration, AI Integration, React, Excel Add-ins, Office Add-ins, TypeScript, System Design, Generative Pre-trained Transformer 4 (GPT-4), Large Language Model Operations (LLMOps), OpenAI API, Azure, AI Design, Optical Character Recognition (OCR), AI Consulting, Machine Learning Algorithms, Amazon Machine Learning, Fine-tuning, Amazon SageMaker, AI Programming, Visual Studio Code (VS Code), Docker, Agentic AI, RAG Systems, API Integration, Vector Databases, Prediction Markets, AI Agent Orchestration, DevOps, LLM Integration, Streamlit, Machine Learning (ML) APIs, Azure Machine Learning, Azure ML Studio, RAG Architecture, Bots, Workflow Automation & System Integration, AI Architecture, Context Engineering
Prompt Engineering Consultant | Data Scientist
2023 - 2024
Toptal
- Acted as a prompt engineering consultant for Toptal’s core team, collaborating with various product teams to automate pre-sales and sales processes using advanced AI solutions.
- Developed AI-generated success stories and evaluation frameworks for sales calls, enhancing the effectiveness of pre-sales strategies and improving customer engagement.
- Created scripts for sales calls and designed objection-handling mechanisms, leveraging RAGs and various LLMs like Cohere for ranking and OpenAI models for content generation.
- Contributed to over 10 projects, providing tailored, prompt engineering solutions that significantly improved Toptal’s sales and customer interaction strategies.
- Used low-code tools like Dify to create LLM pipelines and advanced code-based frameworks like DSPy to automate the prompting process.
Technologies: Prompt Engineering, Large Language Models (LLMs), Artificial Intelligence (AI), Python, APIs, Retrieval-augmented Generation (RAG), OpenAI, Machine Learning, Generative Pre-trained Transformers (GPT), Generative Pre-trained Transformer 3 (GPT-3), OpenAI GPT-3 API, OpenAI GPT-4 API, Data Science, Claude, Generative Adversarial Networks (GANs), Luigi, Apache Airflow, Dagster, Dify, Anthropic, System Design, Generative Pre-trained Transformer 4 (GPT-4), Large Language Model Operations (LLMOps), OpenAI API, Sentiment Analysis, AI Design, AI Consulting, Machine Learning Algorithms, Amazon SageMaker, AI Programming, Visual Studio Code (VS Code), Agentic AI, Open-source LLMs, API Integration, LLM Integration, Gemini, Agentic AI Systems, AI Product Strategy, Strategy
AI Expert
2023 - 2024
Freede Solutions Inc.
- Created a state-based chatbot for debt collection, navigating through different stages during the debt collection process, including guardrails to avoid unexpected behavior.
- Fine-tuned various LLMs for different generative and multilabel classification tasks for various learning techniques such as fine-tuning, few-shot learning, and zero-shot learning.
- Evaluated more than 15 LLMs of different sizes (7B - 180B parameters) for fixed debt collection scenarios.
- Generated conversational, synthetic data for various tasks.
Technologies: Python, Artificial Intelligence (AI), Machine Learning, Generative Pre-trained Transformers (GPT), OpenAI, Chatbots, Chatbot Conversation Design, Natural Language Processing (NLP), AIOps, RunPod, Serverless, APIs, Deployment, Training, Modeling, AI Chatbots, Time Series Analysis, Transformer Models, Time Series, Generative Artificial Intelligence (GenAI), PDF, Databases, Cloud, Software Architecture, Models, Data, AI Integration, LangChain, Prompt Engineering, AI Agents, System Design, Large Language Model Operations (LLMOps), Minimum Viable Product (MVP), Copywriting, OpenAI API, Sentiment Analysis, Azure, Machine Learning Algorithms, Fine-tuning, Amazon SageMaker, AI Programming, Visual Studio Code (VS Code), Docker, Unsupervised Learning, Data Engineering, Agentic AI, Hugging Face, Open-source LLMs, Hyperparameter Tuning, AI Tools, Hugging Face Transformers, Machine Learning (ML) APIs, Azure Machine Learning, Azure ML Studio, Model Development, Model Evaluation, Random Forests, Strategy
NLP Engineer
2023 - 2024
Hao Ting Yen (Scoop AI)
- Developed an automated pipeline to scrap news from various internet sources and categorized them among multiple categories. Deployed in on AWS Cloud Instance.
- Generated podcast transcripts using Open AI LLMs with scrapped articles from various sources.
- Implemented text to speech (TTS) feature to generate podcast audio clips using the transcripts generated by OpenAI LLMs.
Technologies: Python, Machine Learning, APIs, OpenAI Gym, Natural Language Processing (NLP), Natural Language Toolkit (NLTK), SpaCy, Artificial Intelligence (AI), Language Models, OpenAI, Custom Models, Generative Artificial Intelligence (GenAI), PDF, Databases, SQL, Cloud, Software Architecture, Models, Data, AI Integration, LangChain, Prompt Engineering, AI Agents, Minimum Viable Product (MVP), Copywriting, Sentiment Analysis, Optical Character Recognition (OCR), AI Programming, Visual Studio Code (VS Code), Docker, Hugging Face Transformers, Whisper, Machine Learning (ML) APIs
Computer Vision Researcher
2023 - 2023
Pixelcut Inc.
- Implemented TryOnDiffusion, the latest paper on virtual try-on from scratch.
- Implemented cross- and self-attention along with a cascaded diffusion pipeline.
- Conducted research on top virtual try-on models and did comparisons for available open-source projects, including CP-VTON, CP-VTON+, and HR-VITON.
Technologies: Computer Vision, TensorFlow, PyTorch, Machine Learning, Artificial Intelligence (AI), Generative Adversarial Networks (GANs), Custom Models, OpenCV, Databases, Cloud, Models, Data, Optical Character Recognition (OCR), Fine-tuning, AI Programming, Visual Studio Code (VS Code), Docker, Text to Image AI, Hugging Face, Data Quality, Prediction Markets, Computer Vision Algorithms, NVIDIA TensorRT, Pytest, Object Detection, Kalman Filtering, Synthetic Data Generation, Scripting, Video Analytics, 3D Pose Estimation
Computer Vision Contractor
2021 - 2022
Agot
- Partnered with the CTO to improve accuracy for detection and tracking methods. Integrated the new pipeline into their system.
- Improved multi-object tracking accuracy for kitchen menu items by 3%. Focused on reducing ID switches per ground truths and tried different weighted cost functions and heuristic-based improvements.
- Increased multi-object detection mAP by 30% by modifying YOLOv5 classification loss to support multi-label object detection.
Technologies: Computer Vision, PyTorch, Machine Learning, Jira, Slack, GitHub, Programming, Amazon Web Services (AWS), Image Recognition, Image Processing, Google Vision API, Image Search, Labeling, PostgreSQL, Recurrent Neural Networks (RNNs), Neural Networks, Convolutional Neural Networks (CNNs), You Only Look Once (YOLO), Quantitative Analysis, FFmpeg, Communication, Generative Adversarial Networks (GANs), Custom Models, OpenCV, Databases, Cloud, Models, Data, Optical Character Recognition (OCR), icr, Fine-tuning, AI Programming, Visual Studio Code (VS Code), Data Engineering, Hugging Face, Data Quality, GitHub Actions, Prediction Markets, AI Automation, Process Automation, Computer Vision Algorithms, DevOps, Edge Computing, NumPy, Object Tracking, Motion Tracking, Object Detection, Kalman Filtering, Synthetic Data Generation, AI Engineering, Video Analytics, Pattern Recognition, 3D Pose Estimation, Motion Capture
Product Development Lead | CPO
2020 - 2022
Adlytic
- Developed audience analytics through CCTV footage for a retail store. Built the pipeline using machine learning and deep learning algorithms with PyTorch, Keras, and TensorFlow in Python.
- Deployed a scalable solution using Docker to process multiple cameras. The product included features such as footfall counting, dwell time, heatmaps generation, and age, gender, and emotion classification.
- Implemented the product's machine learning and deep learning models in PyTorch and TensorFlow and later optimized using the C++-enabled framework, TensorRT.
Technologies: Jira, GitHub, Slack, Management, Computer Vision, Deployment, Machine Learning Operations (MLOps), NVIDIA CUDA, Jetson TX2, Artificial Intelligence (AI), Python, PyTorch, TensorFlow, Deep Learning, Machine Learning, Algorithms, Keras, Docker, Jupyter, Programming, Integration, Amazon Web Services (AWS), Image Recognition, Back-end Development, Neural Networks, Convolutional Neural Networks (CNNs), You Only Look Once (YOLO), Regression Modeling, Quantitative Analysis, Leadership, Communication, Custom Models, OpenCV, Industrial Internet of Things (IIoT), Databases, SQL, Software Architecture, Models, Data, Predictive Analytics, Minimum Viable Product (MVP), AI Consulting, Fine-tuning, Visual Studio Code (VS Code), Data Engineering, Hugging Face, Snowflake, Startups, Data Quality, GitHub Actions, Predictive Modeling, Prediction Markets, Process Automation, Graphics Processing Unit (GPU), NVIDIA TensorRT, DevOps, Facial Recognition, Facial Tracking, CCTV, Video Surveillance, Edge Computing, NumPy, Technical Leadership, Object Detection, Kalman Filtering, AI Product Strategy, Scripting, AI Engineering, Business Intelligence (BI), Video Analytics, Event-driven Architecture, Solution Architecture, AI Architecture, Pattern Recognition, 3D Pose Estimation, Motion Capture, MLflow, Feature Engineering, Model Deployment, Model Development, Monitoring, IT Strategy, Enterprise Architecture, Fractional CTO
AI Lead | CTO
2019 - 2022
Omno Ai
- Spearheaded the development and launch of OMNO AI's flagship products, Adlytic, Trafflytic, and SportsEye, leveraging AI to revolutionize person, traffic, and sports data analytics for videos and live streams.
- Engineered a cutting-edge, scalable traffic analytics solution deployed for the Indonesian and Turkish governments. Features included real-time traffic density analysis, classified directional counting, and precise lane counting capabilities.
- Increased the AdMob engine's conversion by 200% by creating a scalable hybrid AdMob recommendation engine with collaborative and content-based filtering. Processed one million inferences daily, with a 24-hour online learning cycle.
- Delivered 50+ cutting-edge AI projects, expertly leading a team of eight data scientists and DevOps engineers from pre-sales to deployment. Conducted thorough code reviews and provided invaluable management expertise.
Technologies: AIOps, Artificial Intelligence (AI), Computer Vision, Data Science, Google Cloud Platform (GCP), Machine Learning, GitHub, AWS Command Line Interface (CLI), Deep Learning, APIs, Dashboards, Management, Jira, Slack, Deployment, Machine Learning Operations (MLOps), Natural Language Processing (NLP), ChatGPT, Generative Pre-trained Transformers (GPT), Pandas, Financial Modeling, Jupyter, Programming, Integration, Natural Language Toolkit (NLTK), SpaCy, Amazon Web Services (AWS), Image Recognition, Image Processing, Web Scraping, Image Search, Labeling, Amazon Rekognition, PostgreSQL, Recurrent Neural Networks (RNNs), Neural Networks, Convolutional Neural Networks (CNNs), You Only Look Once (YOLO), Open Neural Network Exchange (ONNX), Regression Modeling, Quantitative Analysis, Forecasting, FFmpeg, Leadership, Communication, Generative Adversarial Networks (GANs), Custom Models, OpenCV, Time Series Analysis, Transformer Models, Time Series, Databases, Software Architecture, Algorithms, Models, Data, Rapid Prototyping, CTO, Predictive Analytics, Data Engineering, AI Consulting, Fine-tuning, Visual Studio Code (VS Code), Startups, Data Quality, GitHub Actions, Clustering Algorithms, Predictive Modeling, Prediction Markets, NVIDIA Jetson, C++, AI Automation, Computer Vision Algorithms, Graphics Processing Unit (GPU), Facial Recognition, Facial Tracking, CCTV, Dlib, NumPy, AI Model Training, Full-stack, YOLOv5, Technical Leadership, Object Tracking, Motion Tracking, Object Detection, Kalman Filtering, AI Product Strategy, AI Engineering, Business Intelligence (BI), Video Analytics, Event-driven Architecture, Solution Architecture, AI Architecture, Pattern Recognition, 3D Pose Estimation, Motion Capture, Pose Estimation, Healthcare, MLflow, LSTM, Feature Engineering, Model Development, Stakeholder Management, Random Forests, XGBoost, IT Strategy, Data Analysis, Fractional CTO
Experience
SportsEye
https://www.youtube.com/watch?v=09Hik9FpzFM&t=53sAn advanced computer vision-enabled deep learning product designed for soccer analysis. It processes real-time soccer broadcast matches, providing comprehensive statistical insights with 24 locality and possession metrics. The system is built upon the following components:
• Goal or attempts event detection using CALF and 3D ResNets.
• Field line segmentation using Pix2Pix-based GANs.
• Player and ball detection using the YOLO SSD family.
• Player re-identification and tracking using visual features and Kalman filter-based tracking methods.
• Camera view classification using EfficientNet models.
• Field key points localization using FLANN-based retrieval on Siamese-based features for field lines and classical holography techniques to estimate camera poses, enabling 3D perspective visualization and enhancing virtual game simulations.
• Timer localization and OCR-based recognition using EAST and EasyOCR-based methods.
With its extensive functionalities and cutting-edge technologies, SportsEye revolutionizes soccer analysis by providing real-time insights and precise statistical data.
• Goal or attempts event detection using CALF and 3D ResNets.
• Field line segmentation using Pix2Pix-based GANs.
• Player and ball detection using the YOLO SSD family.
• Player re-identification and tracking using visual features and Kalman filter-based tracking methods.
• Camera view classification using EfficientNet models.
• Field key points localization using FLANN-based retrieval on Siamese-based features for field lines and classical holography techniques to estimate camera poses, enabling 3D perspective visualization and enhancing virtual game simulations.
• Timer localization and OCR-based recognition using EAST and EasyOCR-based methods.
With its extensive functionalities and cutting-edge technologies, SportsEye revolutionizes soccer analysis by providing real-time insights and precise statistical data.
Adlytic
https://adlytic.ai/An AI-powered retail analytics technology that analyzes the store's audience analytics through CCTV footage using detection and tracking models. I built its computer vision pipeline using machine learning and deep learning algorithms in Pything with PyTorch, Keras, and TensorFlow. The solution is deployed using Docker, is scalable to process multiple cameras, and involves the following features:
• Footfall counting
• Dwell time
• Heatmaps generation
• Age, gender, and emotion classification
• Area-wise conversion
• Staff or customer classification
• Intrusion detection
This product's machine learning and deep learning models were implemented in PyTorch and TensorFlow, which were later optimized using the C++-enabled framework, TensorRT.
• Footfall counting
• Dwell time
• Heatmaps generation
• Age, gender, and emotion classification
• Area-wise conversion
• Staff or customer classification
• Intrusion detection
This product's machine learning and deep learning models were implemented in PyTorch and TensorFlow, which were later optimized using the C++-enabled framework, TensorRT.
Smart Gandola
Smart Gandola incorporates an audience detector based on cutting-edge computer vision algorithms. This set up seamlessly captures frames and transmits them to the Cloud-based pipeline deployed on AWS EC2. On the cloud instance, I deployed a sophisticated pipeline that integrates state-of-the-art models, including face detection, mask detection, frontal face classifier, age classification, gender classification, and emotion classification.
Using these powerful algorithms, Smart Gandola generates comprehensive statistics such as total traffic count, dwell time, and gaze time. These metrics provide invaluable insights into consumer behavior, enabling businesses to make data-driven decisions and optimize their advertising strategies.
To enhance personalization, I implemented advanced age and gender profiling techniques within Smart Gandola. By developing a user-friendly dashboard, businesses can now have precise control over advertisements, allowing them to assign specific ads to predefined age and gender groups. This level of personalization ensures that the right message reaches the right audience, maximizing the impact and return on investment for advertising campaigns.
Using these powerful algorithms, Smart Gandola generates comprehensive statistics such as total traffic count, dwell time, and gaze time. These metrics provide invaluable insights into consumer behavior, enabling businesses to make data-driven decisions and optimize their advertising strategies.
To enhance personalization, I implemented advanced age and gender profiling techniques within Smart Gandola. By developing a user-friendly dashboard, businesses can now have precise control over advertisements, allowing them to assign specific ads to predefined age and gender groups. This level of personalization ensures that the right message reaches the right audience, maximizing the impact and return on investment for advertising campaigns.
Ad-mob Recommendation Engine
By combining collaborative and content-based filtering techniques with the knowledge base, this powerful engine ensures the delivery of highly relevant and personalized recommendations.
Implemented using PySpark on AWS EMR, the engine updates the collaborative and content-based filters every 24 hours to provide the best recommendations daily. The FAST Inference API, hosted on EC2, scales efficiently using a Kubernetes cluster with multiple pods to handle over 8 million inferences per day while meeting the efficiency criteria of 300ms per request.
To support efficient inference, a Redis instance is used to cache results. The deployment strategy follows the green-blue approach to minimize downtime and ensure uninterrupted service during updates after each online learning cycle.
An online learning mechanism was set in place using the green-blue deployment strategy. The system was able to improve the conversion of the Ad Mob engine by 200%
Implemented using PySpark on AWS EMR, the engine updates the collaborative and content-based filters every 24 hours to provide the best recommendations daily. The FAST Inference API, hosted on EC2, scales efficiently using a Kubernetes cluster with multiple pods to handle over 8 million inferences per day while meeting the efficiency criteria of 300ms per request.
To support efficient inference, a Redis instance is used to cache results. The deployment strategy follows the green-blue approach to minimize downtime and ensure uninterrupted service during updates after each online learning cycle.
An online learning mechanism was set in place using the green-blue deployment strategy. The system was able to improve the conversion of the Ad Mob engine by 200%
Medical Insurance Billing Advisor GPT-3-based Chatbot
This project involved the fine-tuning of the powerful Divinchi-03 model on the specific guidelines set forth by the American Medical Association (AMA) for the year 2023. Leveraging the OpenAI API, I embarked on a meticulous process that involved the creation of a comprehensive question-and-answer (Q/A) dataset tailored to the AMA's guidelines.
By fine-tuning the Divinchi-03 model using this carefully curated dataset, I ensured that the model possessed a deep understanding of the latest medical practices and recommendations outlined by the AMA. This state-of-the-art model became an invaluable resource for healthcare professionals, offering accurate and up-to-date answers to a wide range of medical queries.
To provide seamless access and integration for end-users, I integrated the fine-tuned model into the client dashboard of True CRM, a prominent medical insurance billing company based in the United States. This integration enabled True CRM's users to directly access the refined model's capabilities, enhancing their efficiency and accuracy in dealing with medical billing and related inquiries.
By fine-tuning the Divinchi-03 model using this carefully curated dataset, I ensured that the model possessed a deep understanding of the latest medical practices and recommendations outlined by the AMA. This state-of-the-art model became an invaluable resource for healthcare professionals, offering accurate and up-to-date answers to a wide range of medical queries.
To provide seamless access and integration for end-users, I integrated the fine-tuned model into the client dashboard of True CRM, a prominent medical insurance billing company based in the United States. This integration enabled True CRM's users to directly access the refined model's capabilities, enhancing their efficiency and accuracy in dealing with medical billing and related inquiries.
Automated Podcast Generation Using OpenAI LLMs
Automated a pipeline with the following features:
• Scrap news data from various Internet sources;
• Categorise data into multiple verticals (Healthcare, Politics, Sports, Tech);
• Summarize news content using OpenAI LLMs;
• Create podcasts using OpenAI LLMs;
• Create audio podcast clips using text-to-speech (TTS).
• Scrap news data from various Internet sources;
• Categorise data into multiple verticals (Healthcare, Politics, Sports, Tech);
• Summarize news content using OpenAI LLMs;
• Create podcasts using OpenAI LLMs;
• Create audio podcast clips using text-to-speech (TTS).
Debt Collection Assistant
https://www.freede.co/In this project, we developed a state-of-the-art chatbot specifically designed for the debt collection industry. This intelligent system is engineered to navigate through the various stages of the debt collection process with high efficiency and sensitivity. The chatbot incorporates advanced guardrails to prevent any unforeseen behaviors, ensuring a smooth and respectful interaction with the clients.
1. State-based navigation
2. Custom-tailored language models
3. Extensive model evaluation
4. Synthetic conversational data generation
The development of this chatbot marks a significant leap forward in automating the debt collection process. By leveraging cutting-edge AI and language models, we have created a system that not only improves the efficiency of debt recovery but also ensures a respectful and constructive engagement with clients. This project represents a blend of technological innovation and practical application, setting a new standard in the field of AI-driven debt collection.
1. State-based navigation
2. Custom-tailored language models
3. Extensive model evaluation
4. Synthetic conversational data generation
The development of this chatbot marks a significant leap forward in automating the debt collection process. By leveraging cutting-edge AI and language models, we have created a system that not only improves the efficiency of debt recovery but also ensures a respectful and constructive engagement with clients. This project represents a blend of technological innovation and practical application, setting a new standard in the field of AI-driven debt collection.
AI Workflow Platform and Agent Runtime for Sales Consulting Automation
https://sbigrowth.com• Built SBI's end-to-end AI workflow platform from scratch, designed to automate the analytical processes their sales consultants previously performed manually. The platform supported three distinct agent architectures—ReAct (tool-calling reasoning loops), Plan-and-Execute (hypothesis-driven multi-step analysis using SBI's proprietary methodology), and Fixed Flows (deterministic automation graphs)—all running on a self-hosted FastAPI back end with MCP tool servers.
• Migrated the entire AI infrastructure from a legacy stack to a production FastAPI-based runtime with multi-model routing across OpenAI, Anthropic, and Gemini APIs. Implemented real-time streaming over SSE and WebSockets, interruptible human-in-the-loop controls, scheduled execution, and dynamic graph specifications enabling drag-and-drop workflow automation for non-technical consultants.
• Built a SkillRouter layer for dynamic prompt routing and delivered automated reporting using SBI-branded templates from live CRM data. The platform reduced manual analyst effort per client engagement and became SBI's production backbone for AI-assisted sales consulting.
• Migrated the entire AI infrastructure from a legacy stack to a production FastAPI-based runtime with multi-model routing across OpenAI, Anthropic, and Gemini APIs. Implemented real-time streaming over SSE and WebSockets, interruptible human-in-the-loop controls, scheduled execution, and dynamic graph specifications enabling drag-and-drop workflow automation for non-technical consultants.
• Built a SkillRouter layer for dynamic prompt routing and delivered automated reporting using SBI-branded templates from live CRM data. The platform reduced manual analyst effort per client engagement and became SBI's production backbone for AI-assisted sales consulting.
Fine-tuned LLM Pipeline for Automotive Valuation
An end-to-end ML and LLM fine-tuning pipeline built for a leading US automotive data analytics company. The system processes 30+ years of historical vehicle transaction data (millions of rows) for next-day auction price prediction and classification. The pipeline includes:
• Data preprocessing and feature engineering using PySpark (rolling aggregations, dynamic windowing, outlier detection, cyclic encodings).
• Time-series forecasting using Temporal Fusion Transformer and LSTM via Darts.
• LLM fine-tuning of Qwen 2.5 (1.5B/7B) using QLoRA via Unsloth and Axolotl.
• Multi-GPU distributed training (up to 4 GPUs) with HuggingFace Accelerate and RoPE scaling up to 128k tokens.
• Structured dataset generation using sliding windows formatted as ShareGPT-style prompts.
• RAG-style caching with embedding models for contextual retrieval.
• Experiment tracking with MLflow on Databricks.
I implemented it in Python with PyTorch on Databricks infrastructure.
• Data preprocessing and feature engineering using PySpark (rolling aggregations, dynamic windowing, outlier detection, cyclic encodings).
• Time-series forecasting using Temporal Fusion Transformer and LSTM via Darts.
• LLM fine-tuning of Qwen 2.5 (1.5B/7B) using QLoRA via Unsloth and Axolotl.
• Multi-GPU distributed training (up to 4 GPUs) with HuggingFace Accelerate and RoPE scaling up to 128k tokens.
• Structured dataset generation using sliding windows formatted as ShareGPT-style prompts.
• RAG-style caching with embedding models for contextual retrieval.
• Experiment tracking with MLflow on Databricks.
I implemented it in Python with PyTorch on Databricks infrastructure.
Education
2023 - 2024
Master's Degree in Artificial Intelligence
Georgia Institute of Technology - Atlanta, Georgia, USA
2015 - 2019
Bachelor's Degree in Computer Science
National University of Computer and Emerging Sciences - Lahore, Punjab, Pakistan
Skills
Libraries/APIs
PyTorch, TensorFlow, OpenCV, Pandas, Natural Language Toolkit (NLTK), SpaCy, Google Vision API, Amazon Rekognition, OpenAI API, REST APIs, Hugging Face Transformers, Dlib, NumPy, LSTM, XGBoost, PySpark, FFmpeg, Stripe, Keras, React, Luigi, vLLM
Tools
GitHub, Slack, Jupyter, You Only Look Once (YOLO), Open Neural Network Exchange (ONNX), NVIDIA Jetson, Whisper, Pytest, n8n, Azure OpenAI Service, AWS Command Line Interface (CLI), Jira, ChatGPT, Claude, Amazon SageMaker, Claude Code, Azure Machine Learning, Azure ML Studio, BigQuery, Jetson TX2, Git, Amazon Elastic MapReduce (EMR), AWS Deployment, OpenAI Gym, Apache Airflow, Zapier, GraphRAG, NX CAD, Retell AI
Languages
Python, SQL, Snowflake, Python 3, JavaScript, TypeScript, C++
Frameworks
LangGraph, Streamlit, Agentic Frameworks, Flask
Paradigms
Rapid Prototyping, Rule-based Programming, DevOps, Real-time Systems, Best Practices, Model Context Protocol (MCP), Business Intelligence (BI), Event-driven Architecture, Management, Synthetic Data Generation, REST, Automation
Platforms
Visual Studio Code (VS Code), Google Cloud Platform (GCP), Docker, Amazon Web Services (AWS), Vertex AI, Vercel, Databricks, Azure Functions, Azure, Harness, Android, NVIDIA CUDA, Amazon EC2, Raspberry Pi, Kubernetes, RunPod, Firebase, Twilio
Storage
PostgreSQL, Databases, Data Pipelines, Google Cloud Storage, Google Cloud, Redis
Industry Expertise
Healthcare
Other
Artificial Intelligence (AI), Software, Data Science, Computer Vision, Machine Learning, Natural Language Processing (NLP), Machine Learning Operations (MLOps), Deep Learning, APIs, Optical Character Recognition (OCR), Video Analysis, Generative Pre-trained Transformers (GPT), Programming, AI Programming, LangChain, Image Recognition, Visualization, Image Processing, Back-end Development, Image Search, Labeling, Recurrent Neural Networks (RNNs), Neural Networks, Convolutional Neural Networks (CNNs), BERT, Regression Modeling, Quantitative Analysis, Forecasting, Research, Communication, Generative Adversarial Networks (GANs), OpenAI, Custom Models, Generative Artificial Intelligence (GenAI), Large Language Models (LLMs), Architecture, AI Chatbots, Time Series Analysis, Transformer Models, Time Series, Retrieval-augmented Generation (RAG), Data Extraction, PDF, Software Architecture, Models, Data, AI Integration, Prompt Engineering, AI Agents, Anthropic, Predictive Analytics, Generative Pre-trained Transformer 4 (GPT-4), Large Language Model Operations (LLMOps), Minimum Viable Product (MVP), Data Engineering, Sentiment Analysis, AI Design, AI Consulting, Machine Learning Algorithms, Amazon Machine Learning, Fine-tuning, API Integration, Multi-agent Systems, Supervised Learning, Agentic AI, Hugging Face, Open-source LLMs, RAG Systems, Benchmarking, Hyperparameter Tuning, Vector Databases, Data Management, Startups, Back-end, Data Quality, Decision Modeling, Agentic RAG Systems, Decision Trees, Clustering Algorithms, Predictive Modeling, Prediction Markets, AI Tools, MLflow, AI Automation, Process Automation, AI Agent Orchestration, Light LLMs, Computer Vision Algorithms, Residual Neural Networks (ResNets), Graphics Processing Unit (GPU), NVIDIA TensorRT, Low Latency, LLM Integration, RAG Pipelines, Audio, Gemini, Facial Recognition, Facial Tracking, CCTV, Video Surveillance, Edge Computing, Linear Regression, Bayesian Statistics, Statistics, Data Labeling, AI Model Training, YOLOv5, Technical Leadership, Object Tracking, Motion Tracking, Object Detection, Kalman Filtering, Machine Learning (ML) APIs, Workflow Automation, Agentic AI Systems, RAG Architecture, Vector Search, AI Assistants, Bots, Workflow Automation & System Integration, AI Product Strategy, Scripting, AI Engineering, Video Analytics, Financial System Implementation, Solution Architecture, LLM Reasoning, AI Architecture, Context Engineering, Fintech, Medical Imaging, Graph Neural Networks (GNNs), Pattern Recognition, Time Series Forecasting, 3D Pose Estimation, Motion Capture, Pose Estimation, HIPAA, Health, Demand Forecasting, Feature Engineering, Model Deployment, Model Development, Model Evaluation, Monitoring, Stakeholder Management, Random Forests, IT Strategy, Enterprise Architecture, Enterprise SaaS, Data Analysis, Agentic Workflow Design, Strategy, AIOps, Algorithms, Generative Pre-trained Transformer 3 (GPT-3), OpenAI GPT-3 API, OpenAI GPT-4 API, Language Models, FastAPI, Recommendation Systems, Financial Modeling, Integration, Web Scraping, Fashion, Text to Image, Leadership, Industrial Internet of Things (IIoT), Cloud, CTO, System Design, Copywriting, icr, LoRa, Mathematics, Unsupervised Learning, Data Quality Governance, Data Governance, GitHub Actions, VoIP, Pgvector, Full-stack, Amazon Bedrock AgentCore, Knowledge Graphs, 3D Reconstruction, DICOM, Dental Care, Neuroscience, Causal Inference, ISO 27001, SOC 2, Government Contracting, Fractional CTO, Dashboards, Deployment, GPU Computing, Big Data, Autoscaling Groups, Chatbots, Videos, Chatbot Conversation Design, Serverless, Training, Modeling, Signal Processing, Text to Image AI, Excel Add-ins, Office Add-ins, Dagster, Dify, Geospatial Data, Data Scraping, Reinforcement Learning from Human Feedback (RLHF), Data Analytics, Full-stack Development, Psychology, Real Estate, Scraping, Data Architecture, Risk Assessment, Underwriting, Credit Underwriting, Finance, QLoRA, Website Data Scraping, Azure Function App, Delta Lake, ReAct Agents, Qwen, ElevenLabs Solutions
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