Biraj
i sometimes build things from scratch to understand how they work, write blogs, play chess (1900 rapid on lichess), badminton, and read books.
github.com/biraj21 
x.com/biraj21_ 
lichess.org/@/biraj21 
linkedin.com/in/biraj21 work experience
(you can click on the role summary to see the details)-
Hanging Piece -
-
founder
building Hanging Piece, an AI chess coach.
-
-
Outspeed -
-
founding engineer
realtime voice AI at Outspeed.
- developed WebRTC based real-time voice AI agent system. used message-driven Actor Model architecture for concurrent units (threads) in the code.
- designed the system to be compatible with OpenAI Realtime API specifications. also built the React SDK to simplify integration for developers.
- added default system tools to our voice AI infra that enable agents to automatically determine when to skip turns or end conversations.
- implemented Voice Activity Detection system using Silero VAD. also implemented semantic VAD, where the model is less likely to interrupt the user.
- built real-time log monitoring infrastructure that ships application logs from cloud to Grafana Loki and streams them to frontend via SSE (server-sent events).
- automated cloud infrastructure deployment by creating Terraform configurations for Nomad cluster on AWS, managing EC2 instances, S3 buckets, Redis ElastiCache, ECR, etc.
- used ClickHouse for high-throughput analytics (~3M+ events/day), ingesting events via async inserts for convenient consumption, and using materialized views for fast aggregated queries.
-
-
LaneSquare Technology Pvt. Ltd. -
-
software developer -
built thinkstack.ai, a multi-tenant B2B AI support platform where businesses can create and deploy production-ready chatbots powered by RAG.
- implemented Retrieval Augmented Generation (RAG) using Pinecone to enable agent to answer questions using users' knowledge base.
- built a user intent detection using tool-calling to route conversations automatically (lead capture, ticket creation, human handoff).
- shipped the full-stack Thinkstack product: GraphQL backend, Flask-based AI services, and Redis-backed caching + message queuing, deployed on AWS (Cognito, S3, API Gateway)
- designed a real-time event notification system (pub/sub) to send backend events (eg. website crawling completion) to the dashboard for reactive UI state changes.
- built multi-tenant access control including sub-users, RBAC roles/permissions, and API key authentication for secure external integrations.
- created Zapier, Slack, WhatsApp, Facebook Messenger and Instagram integrations for Thinkstack.
-
software developer intern -
migrated Pickcel's backend codebase from node 12 to node 18. also learnt Docker.
- worked with TypeScript, Node.js, and Mongoose to update and optimize existing REST APIs by making some things concurrent. set up integration testing CI using GitHub Actions.
- optimized Dockerfile for our on-premise setup by properly reordering layers, reducing consecutive build times and improving cache utilization. also added redis to the stack.
- introduced semantic git commits, coding conventions, and structure, which is now embraced by all developers in the organization.
-
skills
- programming languages: JavaScript, TypeScript, Python, Go, C, Bash
- libraries and frameworks: React, Node.js, Express.js, FastAPI, GraphQL
- infra & tooling: Docker, AWS, Terraform, GitHub Actions, HashiCorp Nomad
- database & analytics: MongoDB, MySQL, Redis, Firebase Firestore, ClickHouse
- auth & payments: Better Auth, Supabase, Firebase, OAuth, Stripe, Razorpay
projects
LLM serving engine from scratch
a toy LLM server built from scratch with batched inference, auto-regressive generation loop for token
streaming, request queueing, and independent sequence completion. supports both streaming and non-streaming
responses via FastAPI. also includes Whisper for STT.
DotDB: in-browser vector database
a lightweight vector database that runs entirely in the browser. supports cosine similarity search,
persistent storage via IndexedDB, and batch operations. uses transformers.js with bge-small-en-v1.5 for generating embeddings. implements brute-force k-NN search with Float32Array for memory efficiency.
try it
Focker: Linux containers
Focker is a toy container runtime written in Go to understand Docker internals. implements Linux namespace
isolation (PID, mount, UTS), runs processes inside an extracted Ubuntu 22.04 root filesystem, and supports
bind mounts for sharing host directories with containers.
Ask My AI: Chrome extension
a Chrome extension that brings AI assistants into a sidebar. highlight any text and press shortcut to get
instant explanations without tab-switching. hit 100+ users and 16+ five-star ratings within days of launch. built it because existing extensions charged for wrapping AI APIs in a new UI, which seemed redundant
when users already pay for ChatGPT/Claude/Gemini subscriptions.
chrome web store
NumPy from scratch in Go
i was implementing Neural Networks from scratch in Go and as a byproduct of it, i created my own version of
NumPy for tensor operations like +, -, *, /, matrix multiplication, transpose etc. it also supports
broadcasting.
TCP server
a single-threaded TCP server written in C with an event loop using the
poll() system call, without any third-party library. the server listens on a port and echoes back
received data. i've also written a simple client to test the server.
Obsy: AI observability platform
Obsy is an AI observability tool that provides insights into AI operations. it has a Node.js SDK that
automatically instruments OpenAI, Pinecone and Vercel's AI SDK. the dashboard shows LLM trace, with timing
and usage metrics for each stage. built this in a hackathon.
Gomon: nodemon for Go files
a CLI tool to run Go programs in watch mode. to watch for changes, it uses kqueue() system call on
macOS and inotify on Linux. no 3rd-party dependencies.
Other
brainfuck interpreter | JSON parser | nano-like CLI text editor | AI search | web crawler | tensor visualizer (demo) | findREp