Build the Future
with AI/ML Engineer
I design and ship multi-agent systems, RAG pipelines, and production ML — from a browser automation platform that runs a real ReAct loop over Chrome DevTools Protocol, to a federated learning system across 3 institutions, to a computer vision pipeline adapting ad creatives for live client billboards. 21 shipped projects, one recurring habit: instrument the failure, then fix the actual bottleneck.
Explore the full collection of my
design and development work.
Inside My Technical
Core
I build AI systems that work on real infrastructure — not just in notebooks.
I'm an AI/ML Engineer who spends most of my time on multi-agent systems: a browser-automation platform with a hand-rolled ReAct loop and a second LLM acting as judge, a 5-agent URL-to-app factory with a genuine self-healing build loop, a 5-agent vision pipeline that adapts ad creatives to extreme billboard aspect ratios for a live client. On the research side, I've shipped a federated learning system across 3 institutions and a first-of-its-kind Urdu stuttering dataset, submitted to IEEE SLT 2026.
The pattern across almost everything I ship: find the actual bottleneck, not the obvious one — data quality beating architecture choice, model behavior breaking a system before browser mechanics ever do, a negative prompt amplifying the exact thing it told the model to avoid. I write these failures down and design around them.
Download CV ↓
1. Generative AI & LLMs
- LangChain & Agentic Workflows
- LLM Fine-Tuning & Prompt Engineering
- LoRA/QLoRA Optimization
- Hugging Face ecosystem
- RAG & Vector Databases (FAISS)
- Generative Applications
2. Machine Learning & MLOps
- PyTorch & TensorFlow Pipelines
- Computer Vision (OpenCV, segmentation)
- Docker & Kubernetes Deployment
- AWS (EC2, S3) architecture
- MLflow & Airflow orchestration
- PySpark for large-scale data
3. Autonomous AI Agents
- Hand-rolled ReAct loops (observe → think → act → evaluate)
- Actor + Judge dual-LLM architectures
- Browser automation via Chrome DevTools Protocol
- Semantic memory (ChromaDB) & task history
- Tool Use / typed function-calling schemas
- HITL safety gates & loop-breaker design
4. Data Engineering
- Data Ingestion scaling
- Automated ETL Pipelines
- PostgreSQL Database mgmt
- Snowflake & Data Warehousing
- NoSQL optimization
- Data Lake architecture
Datacamp Certified AI Engineer
DataCamp
Deep Learning Specialization
Deeplearning.ai
Generative AI for Everyone
Deeplearning.ai
Generative AI Engineering Specialization
IBM
Machine Learning with Python
IBM