AI/ML Engineer at Jaguar Land Rover, focused on GenAI systems, scalable ML infrastructure, and translate research into deployable products that solve real problems.
Self-healing CI agent that detects failing builds, diagnoses issues from logs, generates fixes using multi-model LLM reasoning, and auto-commits verified solutions. Features hyper-fuzzy patching and deterministic guardrails for 90% reliability.
Full-stack code quality analysis platform combining AST parsing for complexity metrics with Google Gemini for automated refactoring. Handles async analysis via Celery.
GitHub App that automatically reviews pull requests using Llama-3. Analyzes diffs for bugs and security risks, posting inline comments directly to PRs.
Multimodal system transforming images into viral-style tweets. Fine-tuned ViT-GPT2 for captioning with LoRA adapters for style transfer.
Production RAG system for document querying with sliding window chunking, FAISS vector search, and conversational memory.
Precise vehicle localization using Factor Graph Optimization with GTSAM library. Presented technical findings at JLR Tech Symposium.
Open to geeking out on tough technical problems, collab opportunities, and building AI products that actually scale.