I'm Sahil Arate, a Computer Science Engineer who builds full-stack and agentic AI systems — from production-grade multi-agent orchestrators to AI-powered decision engines that actually ship.
I currently work as a Software Engineer at Seratek Systems, and previously interned as a Full Stack Developer at SuperMentr, a Full Stack Web Development Intern at Prodigy Infotech, and a Cloud Security Intern at Eyesec Cyber Security Solutions.
I'm an open source contributor to cal.com and Supabase, and I won 1st Prize at a National Level IoT Hackathon for an end-to-end RFID attendance system.
🎯 Currently: open to full-time SDE / Full-Stack Developer opportunities 🛠️ I work with: Python, TypeScript, Go, React, Next.js, FastAPI, Node.js, NestJS, PostgreSQL, MongoDB, AWS, Docker, Kubernetes 🤖 I like building: agentic AI systems — LangGraph orchestrators, RAG pipelines, knowledge-graph tools, and deterministic decision engines that keep AI honest 📫 Reach out: always happy to connect about roles, collaborations, or open source
Languages
Frontend & Backend
AI & Automation
Databases & Observability
Cloud & DevOps
Autonomous data orchestrator built with LangGraph, coordinating 5 specialized agents over a typed StateGraph to turn plain-English input into structured enterprise schedules. Streaming backend on FastAPI + WebSockets, with semantic context tracking via pgvector and OpenAI embeddings — cut manual adjustments by 40%.
Python FastAPI LangGraph LangChain WebSockets PostgreSQL Redis Next.js Docker
Automated AP system where an LLM handles extraction only — a fully deterministic rule engine independently decides APPROVE/FLAG/REJECT. Matches invoices against POs at both header and line-item level (fuzzy matching via RapidFuzz), with OCR fallback for scanned PDFs and a complete auditable reasoning trail for every decision.
Python FastAPI Pydantic Groq LLM Tesseract OCR RapidFuzz Next.js TypeScript Docker
Analytical intelligence pipeline that maps relational database architecture into visual knowledge graphs, converting plain-text queries into optimized SQL. Dual-database sync between PostgreSQL and Neo4j via SSE streaming, speeding up anomaly discovery by 50%.
Python FastAPI PostgreSQL (Neon) Neo4j AuraDB Cytoscape.js Next.js Docker
Ask questions about a legacy codebase like you were there when it was written. Built for the WeMakeDevs × Cognee Hackathon — ingests source code, git history, and docs into a Cognee knowledge graph so you can query why code was written the way it was.
Next.js TypeScript FastAPI Cognee React Flow Vercel Render
I'm actively looking for SDE / Full-Stack roles and love connecting with fellow builders. Drop me a message I reply fast! 🚀
