Vivek S.

I build production-grade agentic AI systems and LLM pipelines, multi-agent workflows, RAG architectures, and LangGraph-based applications that ship to real users, not just demos. Currently AI Engineer at Multiscale AI, a semiconductor process optimization startup, where I design LangGraph agentic pipelines and architect identity infrastructure on the Ory stack (Kratos, Oathkeeper, Keto). Previously SDE2 at Oracle Financial Services Cloud, where I built high-throughput event-driven microservices on Kafka, Redis, and Spring Boot for ~2.5 years. Recent builds: → Aegis, autonomous multi-agent PR review system. LangGraph orchestration, FAISS-backed RAG over codebases, MCP server integrations, confidence scoring with human-in-the-loop escalation. → LexRAG, legal-domain RAG pipeline. PDF ingestion, hierarchical FAISS indexing, HyDE query expansion, CrossEncoder reranking, RAGAS evaluation harness. → FlatChef, agentic grocery procurement app. LangGraph + Swiggy Instamart MCP + Bhashini ASR + FastAPI + Next.js + Supabase + Telegram Bot. Core stack: LangGraph, LangChain, OpenAI / Anthropic APIs, MCP, FAISS, RAG, Python, FastAPI, Docker, backed by 3+ years of production backend engineering (Java, Spring Boot, Kafka, Redis). Education: B.Tech in Artificial Intelligence, IIT Hyderabad (CPI 8.55, 2023). Certifications: AWS AI Cloud Practitioner, Oracle Generative AI Professional, Oracle AI Vector Search Professional. I'm best for projects that need someone who can take an agent system from "works in a notebook" to "runs in production with evals, retries, observability, and an architecture that holds up." If that's what you need, let's talk