Shaurya Patil

I’m an Artificial Intelligence Engineer focused on building production-ready LLM-powered agentic systems. I specialize in LangGraph/LangChain workflows, RAG pipelines, and fine-tuning LLMs (LoRA/QLoRA) to deliver measurable improvements in accuracy and user experience.

In my OneGreenDiary role, I architected a LangGraph-based agentic AI layer connecting POS and ERP systems for natural-language querying via FastAPI and SQL. I reduced manual retrieval effort by ∼60% and achieved sub-100ms latency using Redis caching, while also building a React/TypeScript + Django + PostgreSQL scheduling platform handling 200+ daily bookings.

Earlier, as an Artificial Intelligence Research Intern at Coding Jr, I led the end-to-end development of an LLM-powered coding copilot deployed to 500+ students across India. By fine-tuning models with LoRA/QLoRA and MLflow-tracked training pipelines, I reduced irrelevant suggestions by ∼35%.

Across my projects, I’ve built a POS agentic AI system with FastAPI tool orchestration and LangSmith monitoring, an ElevenLabs-powered conversational AI call agent with Twilio and Google Calendar scheduling, and a multilingual News RAG platform using mBART-50 + vector retrieval. I’m excited to bring this low-latency, multi-agent orchestration mindset to products that need reliable, explainable AI.