Divesh Bari

I’m a final-semester B.E. (AI & Data Science) engineer specializing in applied AI/ML systems—agentic architectures, LLM integration, and full-stack AI product development. I architected a full-stack autonomous SIEM platform with an LLM-driven multi-agent orchestrator and engineered a forecasting model achieving state-of-the-art accuracy (1.85°C MAE) on a public benchmark.

I build AI systems that work in real workflows: an autonomous LLM-powered triage agent (Llama 3) integrated into a production-style platform, mapping events to a structured knowledge framework (MITRE ATT&CK) using retrieval and reasoning logic. I also co-developed the surrounding FastAPI + PostgreSQL backend and React 18 frontend, including a real, authenticated data pipeline with REST ingestion, API-key auth, and per-user data isolation.

I’m driven by systems that run autonomously and improve detection-to-response loops. In a multi-agent LangGraph setup, I implemented continuous OODA-loop decision cycles with concurrent specialized agents to achieve zero manual intervention across full detection-to-response cycles, and I’ve reinforced this engineering mindset through AI/Data Science internships spanning time-series forecasting and end-to-end ML pipelines.