Full-Stack & AI/ML Engineer | MCA @ Christ University
Former Technical Consultant Intern @ Adobe Consulting Services
Building production agentic workflows, on-device RAG systems, and high-throughput backend architecture.
LinkedIn • Email • Bengaluru, India
- Problem: Enterprise financial planning workflows suffered from slow sequential tool execution and high latency.
- Impact: Reduced planning latency by 60% via async LangGraph graph execution & Pydantic tool schemas.
- Tech: FastAPI, LangGraph, Supabase, PostgreSQL RLS, React 19
🔹 Local-First Agentic RAG — docSeek
- Problem: Data privacy concerns & API costs block enterprise adoption of cloud RAG systems.
- Impact: 100% offline Corrective RAG with sub-15ms search across 10,000+ indexed pages.
- Tech: LangGraph, FAISS (768-dim), SQLite FTS5, Ollama, Kokoro TTS
🔹 Narrative Intelligence Platform — ClearNews
- Problem: Ungrounded LLM summaries obscure news narrative drift over time.
- Impact: Fused fine-tuned BERT bias classification with HDBSCAN/UMAP clustering over 1,000+ daily articles.
- Tech: BERT, XGBoost, SHAP, pgvector, LangGraph, FastAPI, Redis
🔹 Open-Source CLI — teacher-sab (npm)
- Problem: Cross-agent skill installation friction across AI harness ecosystems.
- Impact: Published a zero-dependency CLI package supporting 10 AI harnesses with byte-identical skill distribution.
- Tech: Node.js, npm registry, YAML parser, E2E testing
- AI / ML: LangGraph, RAG/CRAG, BERT, XGBoost, SHAP, FAISS, pgvector, Ollama, PyTorch
- Backend: Python, FastAPI, Node.js, PostgreSQL, Supabase, Redis, Docker
- Frontend: TypeScript, React 19, Tailwind CSS, Vite