Building production AI systems from research and evaluation to APIs, deployment, and real-world use.
I'm a Lead AI Engineer focused on reliable, production-grade systems across LLMs, RAG, AI agents, NLP, and multimodal AI.
- 🧠 Current: Lead AI Engineer at Clinical Pearl
- 🤖 AI Systems: RAG, agent workflows, document intelligence, LLM APIs, automation
- 🔗 Agents: LangGraph, LangChain, tool calling, conditional routing, multi-agent orchestration
- 📚 Retrieval: Chroma, Elasticsearch, Pinecone, embeddings, hybrid search, re-ranking
- 🧩 LLMs: OpenAI, Claude, LLaMA, Mistral, Ollama, Hugging Face
- ⚙️ Engineering: Python, FastAPI, Docker, GitHub Actions, AWS, Azure, SQL, Linux
- 🔬 Research: Vision-language models, medical imaging, low-resource NLP
- 🎓 Education: MS Data Science — FAST-NUCES
- 📍 Based in: Islamabad, Pakistan
|
1,000+ Production Users |
40% Lower Query Latency |
20%+ Better Answer Relevance |
35% Fewer Deployment Failures |
|
Voice-first AI assistant exploring conversational and agentic workflows for healthcare applications.
|
Multi-agent system for clinical research assistance and knowledge-intensive AI workflows.
|
|
Retrieval-augmented question answering over clinical documents using LLM-based generation and document retrieval.
|
Multimodal vision-to-language system for generating radiology reports from chest X-ray images.
|
More selected AI / NLP projects
- 🌐 Urdu → English Neural Machine Translation — Seq2Seq and Transformer-based low-resource translation
- 💬 Rasa Negotiation Chatbot — conversational AI for automated product negotiation
- 🧠 End-to-End QA Bot — deep-learning question answering implementation
- 📄 OCR Text Extraction — OCR and image-processing pipeline
LLMs & Generative AI
OpenAI · Claude · LLaMA · Mistral · Ollama · Hugging Face · LoRA · QLoRA · Prompt Engineering · Evaluation
Agents & Automation
LangGraph · LangChain · Tool Calling · Multi-Agent Orchestration · Conditional Routing · Human-in-the-Loop
RAG & Retrieval
Chroma · Pinecone · Elasticsearch · Embeddings · Hybrid Search · Re-Ranking · Grounded Generation
ML, NLP & Vision
PyTorch · TensorFlow · Scikit-Learn · Transformers · BERT · Vision Transformers · OCR · Medical Imaging
Engineering & MLOps
Python · FastAPI · Docker · GitHub Actions · AWS · Azure · SQL · Git · Linux
Apr 2024 — Present · Remote
- Build and ship production RAG and LLM applications serving 1,000+ users.
- Design Chroma + Elasticsearch hybrid retrieval, embedding, chunking, ingestion, and indexing pipelines.
- Fine-tune and evaluate LLaMA, Mistral, and OpenAI models using LoRA / QLoRA.
- Serve AI capabilities through FastAPI, Docker, async workflows, and CI/CD.
- Lead a team of engineers and collaborate directly with US-based stakeholders.
Jan 2023 — Mar 2024
- Delivered NLP and Generative AI services for question answering, summarization, extraction, and classification.
- Benchmarked LLMs across quality, latency, and cost.
- Built reusable LangChain-based components and integrated AI services through REST APIs.
Earlier teaching experience
- Machine Learning Trainer — NUST SEECS · Jul 2023 — Nov 2023
- Teaching Assistant, NLP — FAST-NUCES · Jan 2023 — Jun 2023
My MS research explored a vision-to-language architecture for automated chest X-ray report generation using a Vision Transformer + GPT-2, with attention-based explainability.
Selected publications
- Optimizing Sentiment Integration in Image Captioning — Computers, Materials & Continua, 2025
- EEBERT: Emoji-Enhanced BERT — IEEE Access, 2024
- Enhancing Emoji-Based Sentiment Classification in Urdu Tweets — IEEE Access, 2024
MS Data Science — FAST-NUCES · 2022–2024
BE Computer Systems Engineering — Mehran University of Engineering & Technology · 2017–2021 · CGPA 3.7/4.0
Certifications: Generative AI with LLMs · AWS Solutions Architect · Machine Learning Specialization · Python for Data Science, AI & Development
Building AI systems that are useful beyond the demo.
Some production, healthcare, and client work is private · Last updated September 2026