AI/ML
Engineer
01
Real-time face recognition, object detection, and video analytics using YOLO, InsightFace, and OpenCV — from live camera integration to production deployment.
02
Multi-agent systems with LangGraph and LangChain, backed by graph and vector databases, that turn business data into plain-English answers.
03
Natural language interfaces to databases, document pipelines, and workflow automation that eliminate manual data entry and reporting.
04
Containerized, cloud-ready deployment of ML services with FastAPI, Docker, and AWS — built for reliability and real-time performance.
Building production AI systems spanning computer vision, LLM agents, and business automation — a face-recognition attendance system (YOLO + InsightFace), a Neo4j/LangGraph business intelligence agent, and natural-language data query tools deployed via FastAPI, Docker, and cloud infrastructure.
Built TransLynx, an end-to-end English-to-Urdu speech translation pipeline with automatic speech recognition (ASR) — trained a CNN on a custom voice corpus using spectrogram and signal processing features, then chained an NLP translation model and a Urdu TTS model into a desktop application.
University of Management and Technology, Lahore — major coursework in Machine Learning, Natural Language Processing, Artificial Intelligence, Data Science, Data Structures & Algorithms, OOP, and Database Management.
Coursera, taught by Andrew Ng / deeplearning.ai — covering neural networks, CNNs, sequence models, and deep learning fundamentals.
Coursera — core AWS services, cloud architecture, and deployment fundamentals for production systems.
University of Management and Technology — coursework certifications covering ML fundamentals, NLP techniques, and exploratory data analysis.
Andrew Ng / Stanford & DeepLearning.AI — supervised learning, regression, classification, and model evaluation fundamentals.
DeepLearning.AI / Coursera — linear algebra foundations underpinning ML models and vector-space methods like cosine similarity.
IBM — EDA, data cleaning, feature engineering, and regression modeling with scikit-learn.
Real-time face recognition and access control system with YOLO + InsightFace, integrated with live RTSP cameras and automated email alerts on unauthorized access.
LangGraph and LangChain agents answering operational and strategic questions in plain English over a Neo4j graph model of business data, with PostgreSQL session memory and a Chainlit frontend.
Natural language agent connected directly to PostgreSQL, cutting data request turnaround from hours to seconds for the sales team.
Automated tool converting Tally CRM Excel exports into Odoo-ready format, replacing hours of weekly manual copy-pasting with a single file upload.
Browser extension that auto-fills Odoo CRM lead fields by extracting contact details from websites on click, eliminating manual data entry.
End-to-end speech translation pipeline with automatic speech recognition (ASR): a CNN trained on a custom voice corpus using spectrogram and signal processing features, chained with an NLP translation model and a Urdu TTS model into a desktop app.
I'm always open to discussing new AI/ML projects, collaborations, or opportunities to build something intelligent together.