Looking for a job

Surajit Sutradhar

@surajitsutradhar

Message

SDE building event-driven systems with Kafka, FastAPI, Redis & AI/LLM integration. Amazon ML Summer School scholar.

India

Message

What I'm looking for

I'm looking to build backend, distributed-systems, and AI-powered products where I can apply event-driven architecture, LLM integration, and machine learning while growing through challenging real-world engineering work.

I'm pursuing Electronics and Instrumentation Engineering at the National Institute of Technology, Silchar, while building production-oriented backend and AI systems.

I built a real-time event-driven recommendation engine using Python, FastAPI, Apache Kafka in KRaft mode, PostgreSQL, Redis, and Docker. I designed weighted scoring, UPSERT persistence, and event deduplication to prevent score inflation from Kafka retry duplicates.

I also developed LPG Inspector, an OpenEnv-based reinforcement-learning environment for industrial inspection decisions, integrating Qwen2.5-72B-Instruct through the OpenAI SDK across 12 deterministic fault profiles. The FastAPI microservice used Docker and WebSocket interaction loops and was deployed on HuggingFace Spaces.

My work spans LoanLens AI, a video-based digital loan origination platform built with Next.js, Supabase PostgreSQL, Clerk, and LLM and speech-processing APIs. I was selected for Amazon ML Summer School 2026 and was a finalist in the Meta PyTorch OpenEnv Hackathon.

Experience

Work history, roles, and key accomplishments

Real-Time Event-Driven Recommendation Engine

GitHub

May 2026 - Jun 2026 (1 month)

Developed a production-grade, event-driven backend processing user events through an Apache Kafka pipeline in KRaft mode, eliminating ZooKeeper to strengthen fault tolerance. Designed a weighted scoring algorithm with UPSERT-based persistence and event deduplication, backed by a Redis cache-aside layer for low-latency reads.