I'm a Generative AI Software Engineer with 4 years of experience building production-grade software at the intersection of LLM systems and backend infrastructure. I've shipped production Generative AI platforms using RAG, semantic and vector search, agentic AI workflows (LangGraph), and Responsible AI guardrails, serving millions of users.
Right now, I'm fine-tuning Qwen3.6-35B-A3B (a 35-billion-parameter large language model) on AMD MI300X GPUs. The model achieved zero refusals across 465 safety tests using a technique called heretic-abliteration, with safety policies enforced at the application layer.
I optimize for latency, cost, and correctness, without the shortcuts that bite you 6 months later.
| Languages |
|
| AI / ML |
|
| Backend |
|
| Frontend |
|
| Databases |
|
| Cloud / DevOps |
|
| Tools |
|
Domain-specialized AI fine-tuning project. Built for the AMD Hackathon 2026. Specialized fine-tune of Qwen3.6-35B-A3B (a 35-billion-parameter large language model) on AMD MI300X GPUs. Achieved zero refusals across 465 safety tests via a technique called heretic-abliteration, with safety policies enforced at the application layer.
Open-source AI research project. Research on safe model abliteration and post-training quantization (GPTQ, AWQ, and GGUF formats). Includes tooling, evaluations, and reproducible recipes for the broader community.
Production-grade agentic AI chatbot. Built with typed tool calls, persistent memory, and streaming responses. Provider-agnostic, so it works seamlessly with multiple AI providers (OpenRouter, Ollama, and Anthropic).
Real-time disease tracking dashboard. Pulls live data from the World Health Organization (WHO) and Johns Hopkins University. Features map-based geospatial visualization and trend forecasting.
I'm currently open to senior or staff engineer roles (individual contributor positions, not management).
I'm looking for opportunities in:
- 🤖 Generative AI / LLM Engineering · RAG, fine-tuning, inference, agentic platforms
- ⚙️ Backend Platforms · high-throughput Python or Java services, distributed systems
- 🛠 Developer Tooling · agentic systems, LLM-powered developer tools
I'm based in India and open to remote roles or relocation. I respond to all messages within 24 hours.
