AI integration developer working in Python and Flutter. I moved into software in 2025 after almost three years as a planning engineer on industrial plant construction.
Open to AI integration, GenAI and Python developer roles in the UAE (ready to relocate, can join immediately) or in Bengaluru and remote.
Unique Fitness Gym: the Flutter front-desk app a gym in Pinarayi, Kerala, uses every day. It handles admissions, payments with PDF receipts, dues and WhatsApp reminders, keeps all its data on the device, and includes a Dart client for the gym's eSSL face-recognition door. 99 automated tests, with CI building web, Android and iOS. Live demo
- ClaimCheck: checks a health insurance claim against the policy wording and IRDAI's rules, works out what should be paid, judges every deduction and drafts the appeal letter. RAG with an agent: hybrid search, LLM tool calling, amounts computed in code and every quote verified word for word. On a labelled test of 20 deductions: 20/20 verdicts, every amount exact, every verdict sourced. Live demo
- Cravio: a Django food-delivery app with Ask Cravio, an AI agent built on LLM tool calling. Five server-side tools search the menus and fill the cart from a plain-English request. 85 tests. Live demo
- PolySub: multilingual subtitles. I benchmarked Qwen3-ASR and Whisper on Google's FLEURS and use the more accurate model per language; LLM translation replies in a fixed JSON schema so every subtitle keeps its timing. 44 tests. Demo
- AI integration: RAG (embeddings, hybrid search, citations), LLM APIs, tool calling, AI agents, MCP, structured outputs, prompt engineering, evaluation on labelled test sets, Whisper, Qwen3-ASR, PyTorch
- Reverse engineering: AI-orchestrated analysis and de-obfuscation of a protected Windows binary with a custom Python harness (Frida, Unicorn, Capstone, pefile)
- Mobile: Flutter, Dart, Hive, Provider
- Backend: Python, FastAPI, Django, REST APIs, MySQL
- Shipping: Git, GitHub Actions, pytest, Flutter tests