I'm a backend engineer. Most of my 11+ years has been on systems where losing or double-handling a message means real money goes the wrong way: payments, banking, and a data platform running 30M+ rows in Postgres. I tend to care most about the boring parts, like catching a bad input and retrying it instead of letting it quietly break something at 2am.

Lately I build with AI on the side too, mostly small RAG and agent projects. A couple are below.

Right now I'm a Senior Software Engineer at Agilent on the manufacturing side, where part of the job is helping the team actually use AI day to day. Before that: HG Insights, Capital One (through Cognizant), and AurionPro.

prateekmulye.dev · LinkedIn · Email

ChatFormula1 — a single-agent agentic RAG assistant for Formula 1 (live) One agent routes between a vector store and live web search, then answers over what it retrieves. Built with production hygiene: auth, rate-limiting, tests, structured logging, caching, graceful degradation, CI/CD, and Pydantic-typed state. LangGraph · Pinecone · Tavily · FastAPI · Pydantic

FinResearch AI — a multi-agent equity-research system (live) A LangGraph manager fans out to specialized research agents in parallel, then an analyst and a reporter write a verdict over what was retrieved. Every step emits structured JSON so the agents stay grounded instead of guessing. My submission to the SuperDataScience CP044 community project. multi-agent · LangGraph · Pinecone · RAG · Python

slipstream-f1-strategist — an event-driven race-strategy service A reactive API publishes a request to Kafka, a consumer runs a deterministic simulation, and the result is written back exactly once. The public, NDA-free analogue of the financial-grade reliability work: bounded retry with backoff and idempotent processing, so a bad message is retried sensibly and never processed twice. Java 21 · Spring WebFlux · Kafka · PostgreSQL · OpenTelemetry · Testcontainers