- 🔭 I'm currently building multi-agent RAG systems, fine-tuned LLMs, and production AI backends — recent projects include Kubrick (multimodal video RAG with Pixeltable, Whisper, GPT-4o Vision, and FastMCP tools) and a text-to-SQL fine-tune of Qwen2.5-Coder-7B using QLoRA/Unsloth on the Spider dataset
- 🌱 I'm currently deepening my skills in agentic architectures (LangGraph), RAG evaluation, and production LLM serving
- 🏢 I currently work as a Software Engineer & Engineering Lead at Aziro Technologies, leading accessibility (WCAG 2.1/2.2 AA) engineering
- 💼 Actively exploring AI Engineer / Fullstack AI Engineer roles at startups
- 📝 I occasionally write articles on medium.com/@gokul1999
- 💬 Ask me about RAG, LLM fine-tuning (LoRA/QLoRA), multi-agent systems, FastAPI, React/Next.js, and Machine Learning
- 📫 How to reach me: sgokulcse10@gmail.com
- Kubrick — Multimodal video RAG pipeline: Pixeltable, Whisper, GPT-4o Vision, FastMCP tools (
process_video,search_video,clip_extraction), Groq, Opik, FastAPI, Next.js 14 - **Text-to-SQL fine-tuning project using Qwen2.5-Coder-7B (QLoRA/Unsloth) on the Spider dataset, served via Ollama/Docker with a FastAPI backend and React frontend. Model published on HuggingFace
Population and Sample Face Recognition Using Face Verification
