Tariq is available for hire
Hire TariqTariq Shatat
Verified Expert in Engineering
Amazon EKS Developer
Cairo, Cairo Governorate, Egypt
Toptal member since March 11, 2022
Bio
Tariq is an AI and cloud solutions architect with deep experience in agentic systems, LLM platforms, semantic data, and enterprise software modernization. He has led teams delivering multi-tenant AI platforms, multimodal RAG, model fine-tuning, and large-scale migrations from monoliths and AWS Lambda to Kubernetes-based microservices. He combines hands-on engineering with strong architecture and delivery leadership.
Portfolio
Ejaaz
Python, Amazon Bedrock AgentCore, Reranking Models, Pgvector, LangGraph...
Carrier - Residential - Mobile Apps Product Management
Team Leadership, Agentic AI, Artificial Intelligence (AI)...
Carrier - Residential - Mobile Apps Product Management
Docker, Amazon, Amazon Kinesis, AWS IoT, SSL Certificates, Load Balancers...
Experience
- Amazon EKS - 6 years
- Kubernetes - 4 years
- MongoDB - 2 years
- Google Cloud - 2 years
- Agentic AI - 2 years
- LangGraph - 2 years
- Amazon Bedrock AgentCore - 2 years
- Apache Kafka - 1 year
Work Experience
AI Cloud Architect
2025 - 2026
Ejaaz
- Enabled Raki Publisher to produce five curriculum textbooks for the upcoming academic year by delivering an AI-powered authoring platform that converted source material into structured, publish-ready educational content.
- Architected an AI curriculum authoring engine with 36 LangGraph pipelines and approximately 216 workflow nodes, automating the creation of lessons, question banks, assessments, and exam-preparation materials.
- Improved closed-book evaluation accuracy from 51% to 83% by fine-tuning Qwen2.5-14B-Instruct with QLoRA on approximately 97,000 Egyptian Thanaweya Amma question-and-answer examples.
- Preserved native tool-calling in the fine-tuned model, merged production adapters, and exported GGUF and Ollama builds to support private, cost-controlled local inference.
- Built a multimodal RAG platform using LlamaIndex, PostgreSQL, and pgvector, combining OCR, Arabic RTL processing, semantic chunking, filtered retrieval, reranking, and page-level citations.
- Delivered traceable AI-generated content by linking retrieved evidence back to source pages, allowing editors to validate lessons and questions before publication.
- Engineered 36 production-grade agentic pipelines with Celery-based orchestration, live progress tracking, cancellation, resumption, failure recovery, and long-running workflow management.
- Designed multi-provider LLM routing across Gemini, Claude, OpenAI, Groq, Grok, and fine-tuned Qwen models, selecting models by task complexity, modality, quality, latency, and cost.
- Reduced dependency on a single model provider by implementing tiered model fallbacks, image-generation fallbacks, provider-specific adapters, and centralized prompt and inference controls.
- Built Arabic-first AI publishing workflows that generated bilingual, RTL-compatible educational content and exported complete projects to Word, PowerPoint, ZIP, and structured JSON formats.
Technologies: Python, Amazon Bedrock AgentCore, Reranking Models, Pgvector, LangGraph, Large Language Models (LLMs), Retrieval-augmented Generation (RAG), Groq, Gemini API, Celery, React, SCORM, SCORM Compatibility, Fine-tuning, qlora
Lead AI and Agentic Solutions Architect
2025 - 2026
Carrier - Residential - Mobile Apps Product Management
- Led a focused AI, data, and platform team while remaining hands-on in architecture, prototyping, integration design, technical validation, and implementation reviews.
- Built the core architecture for WillisAI, an agentic HVAC diagnostic platform that combined LLM reasoning with telemetry, alerts, fault data, equipment context, and technical knowledge.
- Implemented agent workflows that called specialized tools for fault diagnosis, troubleshooting, replacement-part identification, equipment upgrades, and proactive service recommendations.
- Designed and prototyped LangGraph-based orchestration flows, including agent state management, tool selection, response validation, failure handling, and multi-step reasoning.
- Integrated Model Context Protocol services to expose telemetry, diagnostics, equipment data, and knowledge sources as governed tools available to AI agents.
- Designed the multi-tenant WillisAI platform, including tenant configuration, MCP allow-lists, shared inference, API access, schema validation, service health checks, and tenant isolation.
- Defined and validated dynamic MCP registration patterns, enabling new data and domain services to be added without changing the core agent orchestration layer.
- Implemented observability patterns using OpenTelemetry to trace model calls, tool execution, agent decisions, latency, failures, and end-to-end request flows.
- Designed and prototyped a federated semantic data layer using Snowflake, Neo4j, GraphQL, Brick, and Haystack to provide agents with consistent equipment and telemetry context.
- Worked directly with engineers to review code, troubleshoot integrations, test agent behavior, refine prompts and tool contracts, and move prototypes toward production-ready platform components.
Technologies: Team Leadership, Agentic AI, Artificial Intelligence (AI), Large Language Models (LLMs), Amazon Bedrock AgentCore, LangGraph, Model Context Protocol (MCP), Retrieval-augmented Generation (RAG), Neo4j, Snowflake, GraphQL Federation, Brick Schema, Project Haystack, OpenTelemetry, Multi-tenant SaaS
Cloud Platform Architect
2022 - 2025
Carrier - Residential - Mobile Apps Product Management
- Led a multidisciplinary engineering team responsible for cloud architecture, platform modernization, DevOps, IoT infrastructure, application migration, and production delivery.
- Led the technical strategy and delivery roadmap for migrating serverless and monolithic workloads to Kubernetes-based microservices on Amazon EKS.
- Directed the decomposition of complex applications into independently deployable services with standardized APIs, observability, containerization, and CI/CD pipelines.
- Architected scalable AWS platforms for connected HVAC products using AWS IoT, Amazon Kinesis, Kubernetes, private networking, load balancing, and secure service communication.
- Established an AWS-based MLOps platform using Kubeflow and Kubernetes to standardize model training, deployment, versioning, monitoring, and operational governance.
- Defined multi-cloud architecture patterns across AWS, Azure, and Google Cloud for residential IoT, data processing, analytics, and home-energy initiatives.
- Led the enterprise migration from a monolithic application architecture to independently deployable microservices, defining service boundaries, APIs, ownership models, and delivery standards.
Technologies: Docker, Amazon, Amazon Kinesis, AWS IoT, SSL Certificates, Load Balancers, Amazon Elastic Container Service (ECS), Scalability, Performance, VPN, Proxies, Amazon Virtual Private Cloud (VPC), Security, Kubernetes, Kubeflow, Architecture Governance
Lead SRE Engineer
2021 - 2022
Swvl
- Collaborated with the team to reduce cloud spending by 40%, leveraging cloud best practices in terms of autoscaling, capacity planning, and performance optimization.
- Worked with the team to migrate infrastructure to Kubernetes infrastructure based on GKE and EKS.
- Co-built a serverless GitHub Runner orchestrator to autoscale and provision the machines used for CI pipelines on AWS using Terraform, Lambda, API Gateway, and CloudWatch.
- Provisioned the infrastructure of code for multi-environment on both GCP and AWS using Terraform and Terragrunt.
- Helped the team migrate the CI pipelines of 25+ services to GitHub Actions using AWS App Runner and CD pipelines that use ArgoCD declaratively.
- Cooperated closely with development to migrate from monolith to microservices architecture while using a wide variety of technologies, including Kafka, Elastic Search, MongoDB, and PostgreSQL.
- Contributed to the implementation of the observability stack using Prometheus and Datadog.
Technologies: Google Kubernetes Engine (GKE), Kubernetes, Helm, Terraform, Node.js, Go, Microservices, Apache Kafka, Elastic, Continuous Delivery (CD), Continuous Integration (CI), Databases, Linux, Infrastructure as Code (IaC), Site Reliability Engineering (SRE), Cloud, Container Orchestration, AWS Cloud Architecture, Containers, Amazon S3 (AWS S3), AWS IAM, Google Cloud Platform (GCP), Networking, JavaScript, Containerization, Serverless Architecture, Serverless, Amazon Kinesis, Amazon RDS, Relational Database Services (RDS), GitOps, AWS CloudFormation, Google Cloud SQL, Git, Architecture, GitLab, AWS DevOps, Grafana, Datadog, Prometheus, TypeScript, Cloud Infrastructure, Scalable Web Services, Web Hosting, Web Scalability, Identity & Access Management (IAM), AWS Amplify, GitHub, Data Science, Azure, YAML, GitHub Actions, SSL Certificates, Amazon EC2, Load Balancers, Amazon Elastic Container Service (ECS), Scalability, Performance, VPN, Proxies, Amazon Virtual Private Cloud (VPC), Security
Senior Cloud Engineer
2021 - 2022
Rackspace Technology
- Helped our clients to migrate and set up their infrastructure using AWS and Kubernetes.
- Built a complete SAAS platform using AWS serverless architecture to stream a MATLAB-like desktop application using AWS Lambda, Cognito, AppStream, and DynamoDB.
- Helped our client create a multi-account and multiregional, AWS-based environment using Terraform and Terragrunt.
- Used AWS to build a complete SaaS platform to stream a MATLAB-like desktop application to clients through a web interface without changing the core implementation of the app.
- Mentored and supervised a group of two or three engineers to deliver infrastructure to our clients effectively.
Technologies: Amazon Web Services (AWS), Node.js, Amazon AppStream, AWS Lambda, Kubernetes, Redis, DevOps, CI/CD Pipelines, AWS CodeDeploy, Go, Amazon API Gateway, Vault, Continuous Delivery (CD), Continuous Integration (CI), Databases, Linux, Infrastructure as Code (IaC), Site Reliability Engineering (SRE), Cloud, Container Orchestration, AWS Cloud Architecture, Containers, Amazon S3 (AWS S3), AWS IAM, Google Cloud Platform (GCP), AWS Elastic Beanstalk, Networking, JavaScript, AWS Cloud Development Kit (CDK), Containerization, AWS Fargate, Serverless Architecture, Serverless, Lambda Functions, Amazon DynamoDB, Amazon RDS, Relational Database Services (RDS), GitOps, AWS CloudFormation, AWS IoT Core, Git, Architecture, GitLab, AWS DevOps, Grafana, Prometheus, Cloud Infrastructure, Scalable Web Services, Web Hosting, Web Scalability, Identity & Access Management (IAM), GitHub, Kubeflow, YAML, GitHub Actions, SSL Certificates, Amazon EC2, Load Balancers, Amazon Elastic Container Service (ECS), Scalability, Performance, VPN, Amazon Virtual Private Cloud (VPC)
Technical Lead
2020 - 2022
Hospitalia Ltd
- Designed and implemented the migration to AWS serverless architecture using AWS Lambda, API Gateway, S3, and Amazon CloudFront.
- Led a team of front-end, mobile, and back-end engineers to develop and maintain Hospitalia application using ASP.NET Core, React, React Native, and Node.js.
- Collaborated with business stakeholders and product designers to define the product roadmap and analyze customer feedback.
Technologies: ASP.NET, React, Amazon Web Services (AWS), AWS Lambda, API Gateways, Amazon Route 53, Amazon Aurora, GitLab CI/CD, AWS Amplify, CI/CD Pipelines, Python, Monitoring, Amazon API Gateway, Continuous Delivery (CD), Databases, Linux, Infrastructure as Code (IaC), SQL, Cloud, AWS Cloud Architecture, Amazon S3 (AWS S3), AWS IAM, AWS Elastic Beanstalk, Networking, JavaScript, Containerization, Serverless Architecture, Serverless, Lambda Functions, Amazon DynamoDB, Amazon RDS, Relational Database Services (RDS), GitOps, Node.js, Git, Architecture, Next.js, GitLab, AWS DevOps, Grafana, Prometheus, TypeScript, Cloud Infrastructure, Scalable Web Services, Web Hosting, Web Scalability, GoDaddy, Identity & Access Management (IAM), YAML, SSL Certificates, DigitalOcean, Amazon EC2, Load Balancers, Scalability, Amazon Virtual Private Cloud (VPC)
Solutions Architect
2019 - 2021
RoboGarden
- Worked closely with the CTO to provide optimization plans for the infrastructure performance, cost, availability, and resiliency.
- Led the DevOps team to implement AWS WAA and streamline the development and integration process.
- Reduced the cloud spending by 25% while the customer base grew to 200% by leveraging scheduled scaling and server-less techniques.
- Migrated most of the services into either a Lambda and API gateway or Docker on ECS based on the use case.
- Built the CI/CD pipelines for RoboGarden Services using Jenkins.
Technologies: JavaScript, Amazon Web Services (AWS), AWS Lambda, Amazon EC2, Amazon EKS, Jenkins, PHP, Node.js, NGINX, DevOps, CI/CD Pipelines, Autoscaling, Python, Monitoring, Go, MongoDB Atlas, Databases, Linux, Infrastructure as Code (IaC), Cloud, AWS Cloud Architecture, Containers, Amazon S3 (AWS S3), AWS IAM, Containerization, Serverless Architecture, Serverless, Lambda Functions, Relational Database Services (RDS), AWS IoT Core, Architecture, Next.js, AWS DevOps, Prometheus, TypeScript, Cloud Infrastructure, Scalable Web Services, Web Hosting, Web Scalability, GoDaddy, Webflow, Identity & Access Management (IAM), AWS Amplify, Azure, YAML, DigitalOcean, Load Balancers
Project History
Real-time Structural Health Monitoring
During the renovation of the city building of Project Calgary, we were required to collect stress and strain data out of a distributed network of sensors across the building as part of a real-time-based structural health monitoring. I built a sensor aggregator gateway, data analysis, visualization, and archival and post-processing web portal within the project running time ov over one year.
Hospitalia Infrastructure Modernization
https://www.app.hospitalia.net/Hospitalia is a healthtech based startup in the MENA region. As part of the growth, I have designed, architected, and implemented the migration from a simple VPS setup to AWS-based serverless architecture, including the CI/CD pipeline for the development team.
Migration of a Large Infrastructure to AWS
I played a key role in a large migration to AWS, where we had to move the database and workloads without downtime. We have used EKS, Kustomize, RDS, GitHub Action Vault, and ArgoCD.
My main role was to help build the Kubernetes infrastructure with best practices of security, availability, and performance in mind.
My main role was to help build the Kubernetes infrastructure with best practices of security, availability, and performance in mind.
WillisAI — Agentic HVAC Diagnostics Platform
https://www.carrier.com/commercial/en/us/news/news-article/carrier-introduces-generative-ai-feature-in-abound-to-elevate-building-operations-insights.htmlWillisAI is an agentic AI platform designed to diagnose HVAC equipment issues by combining large language models with telemetry, alerts, fault codes, equipment metadata, troubleshooting procedures, and technical documentation. I led a focused AI and platform team while remaining hands-on in the architecture, prototyping, and integration design.
I designed the LangGraph-based orchestration layer, agent-state management, tool contracts, response validation, and Model Context Protocol integrations to expose diagnostic and knowledge services to the AI. I also shaped the platform’s evolution from a product-specific assistant into a reusable multi-tenant AI service with tenant isolation, shared model inference, controlled tool access, dynamic MCP registration, schema validation, and OpenTelemetry tracing.
The platform supported workflows for fault diagnosis, guided troubleshooting, replacement-part identification, equipment upgrade recommendations, telemetry correlation, and proactive service opportunities.
I designed the LangGraph-based orchestration layer, agent-state management, tool contracts, response validation, and Model Context Protocol integrations to expose diagnostic and knowledge services to the AI. I also shaped the platform’s evolution from a product-specific assistant into a reusable multi-tenant AI service with tenant isolation, shared model inference, controlled tool access, dynamic MCP registration, schema validation, and OpenTelemetry tracing.
The platform supported workflows for fault diagnosis, guided troubleshooting, replacement-part identification, equipment upgrade recommendations, telemetry correlation, and proactive service opportunities.
Ejaz Author — AI Curriculum Authoring Platform
https://ejaaz.online/Description
Ejaz Author is an AI-powered curriculum production platform that transforms textbooks and curriculum documents into structured lessons, question banks, assessments, and print-ready educational materials. I led the AI and cloud architecture and worked hands-on across agentic workflows, model fine-tuning, multimodal retrieval, back-end services, and AWS infrastructure.
I architected 36 LangGraph pipelines comprising approximately 216 workflow nodes, supported by Celery orchestration with live progress, cancellation, resumption, and failure recovery. I built a multimodal RAG system using LlamaIndex, PostgreSQL, pgvector, OCR, Arabic RTL processing, semantic chunking, reranking, filtered retrieval, and page-level citations.
I also fine-tuned Qwen2.5-14B-Instruct with QLoRA on approximately 97,000 Egyptian secondary-school Q&A examples, improving held-out closed-book accuracy from 51% to 83%. The platform was adopted by Raki Publisher to produce five textbooks for the upcoming academic year.
Ejaz Author is an AI-powered curriculum production platform that transforms textbooks and curriculum documents into structured lessons, question banks, assessments, and print-ready educational materials. I led the AI and cloud architecture and worked hands-on across agentic workflows, model fine-tuning, multimodal retrieval, back-end services, and AWS infrastructure.
I architected 36 LangGraph pipelines comprising approximately 216 workflow nodes, supported by Celery orchestration with live progress, cancellation, resumption, and failure recovery. I built a multimodal RAG system using LlamaIndex, PostgreSQL, pgvector, OCR, Arabic RTL processing, semantic chunking, reranking, filtered retrieval, and page-level citations.
I also fine-tuned Qwen2.5-14B-Instruct with QLoRA on approximately 97,000 Egyptian secondary-school Q&A examples, improving held-out closed-book accuracy from 51% to 83%. The platform was adopted by Raki Publisher to produce five textbooks for the upcoming academic year.
Certifications
OCTOBER 2024 - OCTOBER 2027
AWS Solutions Architect - Professional
Amazon Web Services (AWS)
JANUARY 2021 - JANUARY 2024
AWS Certified Solutions Architect - Associate
Amazon Web Services (AWS)
Skills
Libraries/APIs
AWS Amplify, Terragrunt, Node.js, React
Tools
Amazon EKS, Google Kubernetes Engine (GKE), Terraform, Vault, AWS IAM, Amazon Virtual Private Cloud (VPC), Helm, NGINX, AWS CodeDeploy, MongoDB Atlas, AWS Fargate, AWS CloudFormation, GitHub, Git, Grafana, Amazon Elastic Container Service (ECS), VPN, Amazon AppStream, GitLab, GitLab CI/CD, Jenkins, Elastic, AWS Cloud Development Kit (CDK), Amazon Kinesis Data Firehose, Celery
Paradigms
DevOps, Continuous Delivery (CD), Serverless Architecture, Continuous Integration (CI), Microservices, Model Context Protocol (MCP), SCORM
Platforms
Kubernetes, AWS Lambda, Amazon EC2, Amazon Web Services (AWS), Docker, Google Cloud Platform (GCP), Linux, AWS Elastic Beanstalk, Azure, DigitalOcean, Apache Kafka, AWS IoT Core, Webflow, Kubeflow, Amazon, AWS IoT
Storage
Databases, Amazon S3 (AWS S3), Google Cloud, Redis, Amazon DynamoDB, MySQL, MongoDB, Amazon Aurora, Google Cloud SQL, Datadog, Neo4j, PostgreSQL
Languages
Go, Python, JavaScript, SQL, TypeScript, Active Server Pages (ASP), PHP, YAML, Snowflake
Frameworks
ASP.NET Core, ASP.NET, Next.js, LangGraph, Django Ninja
Other
Infrastructure as Code (IaC), Site Reliability Engineering (SRE), Cloud, Container Orchestration, AWS Cloud Architecture, Networking, AWS Certified Solution Architect, Containerization, Serverless, GitOps, AWS DevOps, Cloud Infrastructure, Scalable Web Services, Web Hosting, Web Scalability, Identity & Access Management (IAM), SSL Certificates, Load Balancers, Scalability, Performance, Argo CD, CI/CD Pipelines, Autoscaling, Monitoring, Amazon API Gateway, Containers, Lambda Functions, Amazon RDS, Relational Database Services (RDS), Prometheus, GoDaddy, GitHub Actions, Proxies, API Gateways, Amazon Route 53, Amazon Kinesis, Architecture, Data Science, Generative Artificial Intelligence (GenAI), Migration, Amazon Neptune, Security, Architecture Governance, Team Leadership, Agentic AI, Artificial Intelligence (AI), Large Language Models (LLMs), Amazon Bedrock AgentCore, Retrieval-augmented Generation (RAG), GraphQL Federation, Brick Schema, Project Haystack, OpenTelemetry, Multi-tenant SaaS, Reranking Models, Pgvector, Groq, Gemini API, SCORM Compatibility, Fine-tuning, qlora, LLM Agents, QLoRA, Model Fine tuning, Llama Index, Gemini, ECS
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