Hi, I'm Subhranshu
Backend and AI Engineer specializing in production LLM systems (multi-agent orchestration, RAG, and semantic search) and the high-throughput backend services and data pipelines that run them at scale.

About

I've spent the last two years as a founding engineer at Everstar, starting right out of an internship, and most of what I learned came from owning things end to end rather than being handed specs.

That looked like weekly sessions with my manager digging into what users were actually struggling with, pushing hard on whether a request was a vitamin or a painkiller before writing a line of code. It looked like late nights with our CEO in New York, arguing through product direction. And it looked like being the person who stayed up until 6 AM to make sure a demo went cleanly because the feature was mine, and shipping it correctly, every edge case, local and production, was my job and not someone else's.

Ownership meant nothing without speed. Early-stage startups don't get the luxury of time, so I shipped at a pace I'd never worked at before, driven mostly by how much I cared about what we were solving: nuclear compliance. I'm looking for the next place to build like that.

Work Experience

Skills

AI & Agents

Multi-Agent Systems
LangGraph
LangChain
LangSmith
Agentic RAG
Hybrid Search
Reranking
LLM Evaluation
Anthropic
OpenAI

Backend & Data

Python
FastAPI
Streaming / SSE
Airflow
Celery
Kafka
OpenSearch
PostgreSQL
MongoDB
Neo4j
Redis

Infrastructure & DevOps

Kubernetes
Docker
Terraform
ArgoCD
GitHub Actions
Helm
GCP
Azure

Observability & Testing

OpenTelemetry
Prometheus
Grafana
Sentry
Pytest
Playwright

Frontend

TypeScript
Next.js
React
Node.js
TailwindCSS
React Query
My Projects

Check out my latest work

I've worked on a variety of projects, from simple websites to complex web applications. Here are a few of my favorites.

Wizz AI

Your knowledge base. Production ready. Wizz AI manages the documents, URLs, and configuration behind your AI chatbot. Add sources, review logs, and tune responses, all from one dashboard.

Next.js
FastAPI
Redis
Celery
Crawl4AI
HuggingFace

Smart Home MCP Server

An MCP server that connects Claude to my homelab, letting it control smart home devices through Home Assistant built with FastMCP and deployed remotely behind an NGINX proxy.

FastMCP
MCP
Python
GenAI
NGINX

Sentiment Analyzer

A sentiment analysis and review management tool that uses Hugging Face's transformers library to analyze the sentiment of given movie reviews.

Next.js
Shadcn UI
TailwindCSS
Hugging Face

RAG Ollama

An Adaptive RAG app that uses Ollama & LangGraph to retrieve answers to user questions from provided documents or from the internet.

Next.js
Langchain
LangSmith
LangGraph
ChromaDB
FastAPI

RudeGPT

An unconventional ChatGPT bootleg with an attitude, achieved by using prompt engineering and a customized prompt engine.

Next.js
AWS Amplify
AWS Cognito
AWS Appsync
OpenAI
Contact

Get in Touch

GitHub
LinkedIn
youtube