Lead AI Engineer and Technical Lead with 14+ years of enterprise software engineering discipline (React, Node.js, TypeScript) and 5+ years of Azure Cloud architecture experience, transitioning into AI Engineering. I specialize in bridging full-stack cloud microservices with production-grade, autonomous AI systems.
Currently architecting secure, autonomous AI workflows using LangGraph, CrewAI, FastMCP (Model Context Protocol), Microsoft AI Foundry, and advanced evaluation frameworks with LangSmith.
⚬ AI Orchestration & Multi-Agent Systems: LangGraph, CrewAI, Model Context Protocol (MCP / FastMCP), Microsoft AI Foundry, Azure AI Agent Service, ReAct, Human-in-the-Loop (HITL), Deep Agents.
⚬ RAG & Vector Architectures: Single-Index Metadata RAG, Chroma DB (HNSW Isolation), Pinecone, FAISS, Semantic Vector Caching, Local Embeddings, Docling PDF Parsing.
⚬ LLMOps, Evals & Guardrails: LangSmith (Tracing & Trajectory Tracking), LLM-as-a-Judge, OpenEvals, Azure AI Content Safety, PII Masking & Redaction Middleware, Pydantic Structured Outputs.
⚬ Backend, Web & Cloud: Python, FastMCP, FastAPI, Node.js, React, TypeScript, PostgreSQL (State Persistence), SQLite3, Streamlit, Gradio, Microsoft Azure (APIM, App Services), Docker.
🤖 LangGraph + CrewAI + MCP Orchestrator
⚬ LangGraph • CrewAI • FastMCP • Python
⚬ Production-grade monorepo demonstrating multi-agent orchestration using LangGraph as a deterministic control plane, CrewAI for reasoning, and FastMCP to expose agentic tools.
📑 AI Resume Analyser & Proof-of-Work Verification
⚬ LangGraph • Streamlit • Pinecone • MCP • LangSmith • Pydantic
⚬ AI-powered candidate evaluation app featuring real-time GitHub commit and repository verification via a custom Model Context Protocol (MCP) server, PII detection, and full observability.
📊 Agentic Investment Insights Engine
⚬ CrewAI • LangGraph • Python
⚬ Multi-agent financial intelligence system that ingests SEC 10-K filings, research notes, and client profiles to generate verified investment insight packs.
🔬 LangSmith RAG Observability & Evaluation RAG
⚬ LangSmith • OpenEvals • LLM-as-a-Judge • Tracing
⚬ Dataset evaluation and trajectory tracking pipeline for benchmarking RAG accuracy, latency metrics, and automated assertion scoring.
🌐 Multi-Entity Single-Index Metadata RAG
⚬ Chroma DB • LangChain • HNSW • Claude 3 • System Design
⚬ Scalable, production-grade RAG architecture querying multi-country datasets using HNSW metadata isolation in Chroma DB.
🚗 UK Driving Theory & Learner Practice iOS App
Swift•iOS•OpenCode•Claude Code- Mobile application designed for UK driving learners to practice hazard perception and theory tests. Built using OpenCode and Claude Code as primary AI coding agents to accelerate Swift/iOS development, setup UI components, and structure mock test engines.
⚬ 🌍 Based in: London, United Kingdom
⚬ 💼 LinkedIn: https://www.linkedin.com/in/vakacharla-sathya/
⚬ ✉️ Contact: svakacharla@yahoo.com
⚬ 🧠 Currently Building: Autonomous Multi-Agent Systems, Custom MCP Servers, and Guardrailed Production RAG Pipelines
