I’m Pritam, a software developer focused on architecture, reliability, and developer experience. I love turning messy real-world constraints into clean systems, observability you can trust, and tooling that teams actually use.
A quick snapshot of how I work, what I value, and what I build.
I like clarity: clear APIs, clear logs, clear ownership, clear failure modes. My happy place is designing systems that survive real traffic, real outages, real humans. I also care about craftsmanship, because the cleanest architecture is the one people can actually maintain.
Stuff I reach for when things need to be fast, reliable, and boring in production.
Designed and evolved a production scheduling system with reliable execution, retries, and audit-friendly logs. Built for scale and real-world failure modes.
Lightweight iOS utility for adding clean, print-ready borders while preserving metadata. Built for photographers who want speed, not a 20-step workflow.
Experimented with local and cloud-hosted LLM serving, RAG pipelines, and developer workflows to generate tests, explain code, and automate ops tasks.
Practical patterns for building boring reliability: jittered backoff, circuit breakers, dead-lettering, and observability conventions.
A structured decision system that scores choices based on alignment, compounding benefits, and artifact output. Designed to reduce indecision and keep momentum.
Migration runbooks, environment-safe configurations, and deployment strategies that minimize downtime and maximize rollback confidence.
Short docs, long thoughts, and practical engineering writeups.
A practical guide to jitter, backoff, idempotency keys, and failure budgets. Includes “what not to do” patterns that cause thundering herds.
Swimlanes, boundaries, responsibilities, and why “one mega diagram” fails. Includes a template that works for enterprise infra discussions.
Fastest way: email. This form opens your mail app (no backend needed).