Umaan
Banjara
Building the parts of an app you never see — until they break. Auth, caches, queues, the plumbing underneath.
Currently deep in Go: authentication, PostgreSQL, Redis-backed caching and rate limiting, background workers, and Docker deployments. Came in through Flutter, stayed for the backend. 18+ public repos, mostly systems built to understand how the real thing works, not just to ship a demo.
Where the work happens
Flutter Developer → Client Solutions
Pi Innovations
- Built and shipped features for the company's Flutter mobile applications before shifting focus toward backend systems and Go.
- Now leads client-facing product demos and requirement-gathering calls, translating customer needs into technical specs for the engineering team.
- Keeps building backend systems on the side — self-directed learning applied directly to production-style personal projects.
Systems, not just scripts
Self-hosted HLS video streaming backend — authenticated upload, background ffmpeg transcoding into chunked segments, and manifest serving with path-traversal protection. JWT auth, Redis-backed caching, and a distributed rate limiter that holds up correctly across load-balanced instances.
Concurrent URL-checking worker pool using goroutines and channels, with per-request timeouts via context cancellation. Fixed a real distributed-systems bug where in-memory rate limiting under-enforced across instances — resolved with a Redis-backed shared counter, proven under an nginx load balancer.
A 1xBet-style sportsbook backend built to learn how the real pieces fit together — wallet locking, odds snapshots, and payout settlement. Live football data synced on a schedule; two background workers handle syncing and bet settlement independently of the request cycle.
A live, browser-based Linux learning environment — hosts run real commands in a sandboxed Docker executor while participants watch, chat, and take quizzes in real time. JWT auth with refresh tokens, WebSockets for live sessions, PostgreSQL underneath.
A Telegram reminder bot with a background scheduler that checks for due reminders every 30 seconds. Redis-backed rate limiting on the API, PostgreSQL for persistence, everything running through Docker Compose as separate bot and API services.
13+ more repos on GitHub — smaller experiments, learning projects, and things still in progress.
see all repos ↗What I build with
- Go
- Python (FastAPI)
- Dart (Flutter)
- SQL
- PostgreSQL
- Redis
- Docker & Compose
- nginx
- Kafka
- ngrok
- JWT auth
- Goroutines & worker pools
- Distributed rate limiting
- Pub/sub, HLS streaming
The path here
Building something? Let's talk.
Open to backend roles and interesting problems — Go, distributed systems, or anything that needs the plumbing done right.
↓ Download résumé