Sathish Raja
Senior Full-Stack Engineer, Building Production AI Systems
14+ years shipping full-stack platforms across government, finance, and robotics in Singapore. Now building multi-agent AI systems with LangGraph, RAG, and MCP — most recently an AI hiring platform built solo, end to end.
Cut deployment errors 50% and lifted API throughput 40% on production robotics platforms at Konica Minolta · 10+ production apps shipped across 4 industries
📦 Featured Projects
Pick a project below to see what it does, how it's built, and the impact it had.
SmartHireAI Platform
AI hiring platform that screens candidates automatically
// Project Scope Summary
A multi-tenant AI hiring platform I designed and built solo, with separate dashboards for candidates, recruiters, and hiring managers. Autonomous AI agents parse incoming resumes and rank candidates against role requirements, removing the manual first-pass screen recruiters normally do by hand.
// Integrated Stack Badges
Data Ingestion & Integration Topologies
Explicit engineering pathways mapped directly out of production system parameters.
LangGraph multi-agent workflow
Qdrant vector embeddings
BullMQ / Redis background jobs
pdf2json resume parsing pipeline
Agent workflows are coordinated with LangGraph and Ollama, so screening runs as a sequence of independent, reviewable steps rather than one opaque call. Resumes are parsed with pdf2json and embedded into a local Qdrant vector database for semantic candidate-to-role matching. Parsing and ranking run as background jobs on BullMQ and Redis, so the app stays responsive while AI work happens behind the scenes. Agents are also wired through the Model Context Protocol (MCP) to safely read file states, check calendars, and send interview invites on their own.