Bridging software engineering and modern AI.
I'm an AI engineer and full-stack developer with 5+ years of production software experience and an MS in Artificial Intelligence from LUMS.
I specialize in LLM applications — RAG systems, agentic AI, and ML/NLP pipelines — and the production backends and web apps around them. I've built government-scale AI services, RAG platforms, LLM evaluation systems, and edge-optimized model deployments.
I care about shipping: production-grade, maintainable systems, communicated clearly. I bridge rigorous software engineering with modern AI.
- Based in
- Lahore, Pakistan
- Education
- MS Artificial Intelligence, LUMS
- Experience
- 5+ years in production
- Focus
- AI / RAG · Full-stack
Where I've worked.
Senior AI Developer
Phaedra Solutions
- Designed and built AskTT, a production FastAPI RAG backend for the MyTT government portal — hybrid retrieval (BM25 + dense kNN + Reciprocal Rank Fusion in Elasticsearch), cited OpenAI answers, KB-only/general modes, and context-aware query rewriting.
- Shipped streaming chat over SSE, AI-assisted search, and Whisper voice chat sharing conversation state across text and voice — on a single backend service.
- Built an LLM recruitment-matching engine evaluating 400,000+ applicant–vacancy pairs with explainable scoring, reaching 30–50 matches/sec while cutting token cost and latency.
- FastAPI
- Elasticsearch
- OpenAI
- Whisper
- PostgreSQL
- Prisma
Senior Software Engineer
Esketchers
- Built an internal Django RAG proposal generator (pgvector, hybrid semantic + keyword search, multi-format docs) on AWS with Nginx/Gunicorn and CI/CD — reaching 90%+ match relevance.
- Developed and maintained Mentalhealth Match (Next.js 14 + Angular SSR, Node/Express, PostgreSQL/Prisma, Redis/BullMQ, Cloudflare) supporting 30,000+ therapists, with 85+ PageSpeed on key pages.
- Contributed the Django/DRF backend and Angular frontend for Solvpath, including a GrapesJS web builder with custom reusable blocks and Pusher real-time notifications.
- Next.js
- Django
- Node.js
- PostgreSQL
- Redis/BullMQ
- AWS
Software Engineer
i2c Inc.
- Supported production digital-payment systems for Visa, Mastercard, Discover, and UnionPay integrations — analyzing APIs, Informix SQL, ETL jobs, and transaction flows.
- Certified payment APIs (Postman, SOAP UI), authored test plans, and validated releases with E2E, smoke, negative, and beta testing in PCI-domain systems.
- Payments
- Informix SQL
- Postman
- SOAP UI
Tools I work with.
A working toolkit across AI/ML, backend, data & search, cloud, and frontend.
AI / ML
- Machine Learning
- Deep Learning
- NLP
- RAG
- Agentic AI
- ReAct
- Prompt Engineering
- LLM Evaluation
- MCP
- Efficient LLMs
- Diffusion Models
- Flow Matching
Frameworks & Models
- PyTorch
- Hugging Face
- LangGraph
- LangChain
- MLflow
- LangSmith
- scikit-learn
- Sentence-Transformers
- Whisper / WhisperX
- FAISS
- ChromaDB
- llama.cpp
- Mistral
- LLaMA
- RoBERTa
- HuBERT
Backend
- Python
- FastAPI
- Django
- Node.js
- Express
- NestJS
- TypeScript
- REST APIs
- SSE Streaming
- Webhooks
- Event-Driven Systems
Data & Search
- PostgreSQL
- pgvector
- Elasticsearch
- BM25
- Dense kNN
- Reciprocal Rank Fusion
- Prisma
- Redis
- BullMQ
- Embeddings
Cloud & DevOps
- AWS (EC2, RDS, S3, Lambda)
- CloudFront
- IAM
- Secrets Manager
- Docker
- GitHub Actions
- Nginx
- Gunicorn
- Cloudflare
- GCP
- Sentry
Frontend
- Next.js
- React
- Angular
- Redux
- Tailwind CSS
- Chakra UI
- HTML5
- CSS3
Academic background.
MS in Artificial Intelligence
Lahore University of Management Sciences (LUMS)
CGPA 3.90 / 4.00
Focus: LLMs, deep learning, computer vision, NLP, efficient AI, agentic AI systems, and AI alignment.
BSc Electrical Engineering
University of Engineering & Technology (UET) Lahore
CGPA 3.595 / 4.00
Final-year project: a single-sensor Non-Intrusive Load Monitoring (NILM) system for appliance-level energy disaggregation.
A-Level
Beaconhouse School System
3 A*s
O-Level
Beaconhouse School System
3 A*s, 4 As, 1 B
Certifications & awards.
Certifications
- Retrieval Augmented Generation (RAG)DeepLearning.AI · Aug 2025
- Neural Networks and Deep LearningDeepLearning.AI · Jul 2025
- AI For EveryoneDeepLearning.AI · Jul 2025
- HTML, CSS & JavaScript for Web DevelopersJohns Hopkins University · Nov 2021
Awards & honors
- Merit-based academic scholarship (two semesters)UET Lahore · 2017–2018
- Laptop AwardChief Minister Punjab · 2018
- Gold medal — 3 A*s at A-Level (nationwide criteria)Beaconhouse · 2016
- Merit-based academic scholarship (A-Level)Beaconhouse · 2014–2016
- Silver medal — 3 A*s & 4 As at O-LevelBeaconhouse · 2014
Ways to work together.
Scope-based engagements — I'll recommend the right fit after a quick chat.
AI / RAG prototype
A focused proof-of-concept — retrieval over your data, a chatbot, or an agent — scoped to validate the idea quickly.
Full-stack MVP or website
From idea to deployed product: a fast, SEO-friendly site or a complete MVP build.
AI chatbot over your knowledge base
A grounded assistant over your docs and data, with citations and sensible guardrails.
Backend & API development
Scalable services, integrations, and data pipelines (FastAPI / Node, AWS).
Maintenance & consulting
Ongoing support and improvements, or advisory on AI and architecture.
From idea to production.
Discovery
Understand the goal, constraints, and what success looks like.
Technical plan
Scope, architecture, and a clear path to ship.
Build
Iterative development with regular check-ins.
Review
Test, refine, and validate against the goal.
Deploy
Ship to production with monitoring in place.
Maintain
Support, iterate, and improve over time.