Hello I'm
End-to-end RAG pipelines and agentic AI workflows with LangChain and LangGraph — orchestration, tracing, and evaluation included.
Embedded directly with enterprise clients to take AI solutions from prototype to production, on cloud or in-house servers.
Model serving, observability, CI/CD automation, and AI platform governance across Azure and AWS.
12% revenue increase from pairing models at Afiniti, 15% forecasting accuracy gains for Estée Lauder, 40,000+ users reached.
I build AI that ships. From national-scale RAG systems to AdTech automation — I embed with teams, take prototypes to production, and leave the infrastructure running.
AdTech · Automation
Problem: Active Amazon ad campaigns needed continuous optimization that manual workflows couldn't keep up with.
Built: A proof-of-concept campaign management platform with backend infrastructure for high-frequency data ingestion and real-time bid adjustments.
Impact: Demonstrated automated campaign operations end-to-end and anchored Utopia Brands' AdTech automation roadmap.
Technologies: Python, AWS, Amazon Ads API
GenAI · Knowledge Graphs
Problem: The Ministry needed reliable, grounded answers over large tourism knowledge bases for strategy and operations.
Built: An enterprise-grade Retrieval-Augmented Generation system with a ReAct agent chatbot — LangChain and LangGraph for orchestration, tracing, and evaluation, with Neo4j Knowledge Graphs and Azure AI Foundry powering the retrieval pipeline.
Impact: Grounded, traceable answers informing strategy and operations at a national ministry.
Technologies: LangChain, LangGraph, Neo4j, Azure AI Foundry
Computer Vision · Manufacturing
Problem: Manual quality assurance on manufactured cloth was slow and inconsistent.
Built: Reverse engineered a fabric inspection machine and integrated computer vision for automated defect detection in production workflows.
Impact: Automated QA on manufactured cloth, embedded directly into the production line.
Technologies: Computer Vision, Python, Embedded
Computer Vision · MLOps
Problem: Enterprise applications needed secure, accurate facial recognition with production-grade availability.
Built: AI-powered FastAPI microservices using Vision-Language Models and Face Recognition, deployed with high availability and real-time inference.
Impact: Serves 1,000+ internal users in production.
Technologies: VLMs, FastAPI, Docker, Azure
Languages & Frameworks: Python, SQL, PyTorch, TensorFlow, LangChain, LangGraph, LangSmith, FastAPI, REST APIs
AI & ML: Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), Agentic AI, Agentic Workflows, Model Context Protocol (MCP), AI Development Kit (ADK), Embeddings, Vector Search, Computer Vision, NLP, Predictive Modeling, Time-Series Forecasting
Cloud & Infrastructure: Azure AI Foundry, Azure ML, Azure Databricks, Azure Cognitive Search, AWS, Docker, Model Serving, Inference Pipelines, Batch & Real-Time Inference, Scalable Infrastructure
MLOps & DevOps: MLOps, LLMOps, CI/CD Automation, Azure DevOps, Git, Observability & Monitoring, GenAI Pipeline Governance, Cost Optimization
Data & Visualization: Neo4j, Knowledge Graphs, Power BI
Domain: HealthTech, AdTech (Amazon Ads API), Financial Forecasting, Computer Vision
Autonomous agents and multi-step agentic workflows built with LangChain and LangGraph — with the orchestration, tracing, and evaluation to run them in production.
Production deployment of Large Language Models and Retrieval-Augmented Generation systems grounded in your data, from vector search to knowledge graphs.
End-to-end ML pipelines on Azure and AWS — CI/CD automation, model serving, observability, and AI platform governance.
Deep-learning vision systems — from automated industrial inspection to face recognition and Vision-Language Model applications.
Agentic AI development with LangChain and LangGraph, LLM deployment and RAG engineering, MLOps/LLMOps pipeline implementation, computer vision solutions, and end-to-end machine learning systems for enterprises — from prototype to production.
Yes. Ahmed led the engineering of an enterprise-grade RAG system with a ReAct agent chatbot for the Ministry of Tourism Saudi Arabia, built with LangChain, LangGraph, Neo4j knowledge graphs, and Azure AI Foundry.
Yes. Ahmed is Microsoft-certified five times over (Azure AI Engineer, Azure Data Scientist, AI Fundamentals, Fabric Analytics Engineer, Fabric Data Engineer) and has shipped production AI systems on both Azure and AWS.
As Director of AI at Utopia Brands he led enterprise-wide GenAI automation, including agentic workflows and AI-powered internal tooling built with LangChain and LangGraph, plus GEO/AEO initiatives for AI-driven search visibility.
Globally, with a presence in Toronto, Canada (+1 647 242-9068) and Karachi, Pakistan (+92 333 3145702). He works with clients across North America, the Middle East, and Asia.
AdTech (Amazon campaign management), tourism (Ministry of Tourism Saudi Arabia), beauty and retail (Estée Lauder), telecommunications (Afiniti), manufacturing (fabric inspection), HealthTech, and education technology (Stemers robotics).
Email map.ai@outlook.com, call +1 (647) 242-9068 (Canada) or +92 333 3145702 (Pakistan), use the contact form on this site, or connect on LinkedIn at linkedin.com/in/parachaai.
Open to consulting engagements and senior AI roles across North America, the Middle East, and Asia.
Email map.ai@outlook.com · +1 (647) 242-9068 (CA) · +92 333 3145702 (PK)