A short reflection on Agency, Consistency, and Ego — a simple framework that’s helped me shape goals and momentum, and might do the same for you.
Anindya Chowdhury
Software Engineer | Scaling Systems & Touching Grass (Sometimes)
নমস্কার / Namaste 🌍
Based in Kolkata 🇮🇳. I build backend systems for a living and spend weekends tinkering with side projects. When I’m not coding, I’m usually watching anime, reading manhwas, losing Valorant matches 🎮, or wandering through history and politics rabbit holes.
Writing
View all postsStep-by-step guide to deploying a Golang backend as an Azure Function using a custom runtime.
Projects
View all projectsSamvad News Harvester is a tiny Go service that fetches news links from provider feeds, enriches them with lightweight metadata, and fans structured events out to queues or webhooks. It is intentionally small, pluggable, and welcoming to new contributors. The harvester only uses public sitemaps, or link lists, and never fetches or stores full article bodies.
Indiclang Fraud Shield is a lightweight, multilingual scam detection assistant for WhatsApp and Telegram. Built for AI in Action Hackathon.
Experience
View full résumé
- Architected and scaled Partner API Service enabling broadcast and media partners to ingest live streams and retrieve real-time metadata; built a metadata orchestration layer to normalize, transform, and deliver AI-generated data in partner-specific formats, supporting production-critical workflows, handling 150K+ events/day at peak, and accelerating partner onboarding. - Implemented external integrations within the Publisher Service (Dropbox, OneDrive, Slack, etc.), automating clip delivery and reducing turnaround time for enterprise media workflows. - Designed the architecture for the match-schedule ingestion pipeline (push + pull modes), defining data flows and failure handling to support rapid onboarding at scale, covering 100+ tournaments, 1,000+ teams, and 15K+ players. - Owned P0/P1 production incidents across partner-facing systems, leading debugging, root-cause analysis, and postmortems; introduced structured on-call and RCA practices that reduced recurrence of partner-visible issues. - Led organization-wide logs, alerts, and monitoring migration from Datadog to Last9 with PagerDuty integration, designing PromQL-based indicators and alert routing to improve incident detection and response across revenue-critical services. - Designed and implemented a dual-write migration strategy to propagate collection-level data changes from MongoDB to PostgreSQL using SNS, achieving <100ms propagation latency, and built backfill pipelines to migrate historical data safely. - Stack: NestJS, PostgreSQL, Redis, SNS, SQS, Lambda, MongoDB.
- Built a webhook-based calendar synchronization system for Google and Outlook, enabling real-time meeting tracking and delivery of AI-generated summaries ; integrated with Google Docs, ClickUp, Notion, and HubSpot to publish post-call summaries, contributing to a ~25% increase in user engagement. - Designed and implemented a Telegram-based AI voice chatbot using LLMs (Pygmalion-7B, Mythallion-13B) and ElevenLabs, handling voice and message delivery via AWS ELB and Lambda. - Stack: Python (Django), MariaDB, Redis, Supabase, AWS Lambda, EBS, SQS, React.js, Redux, Slate.js
Reading List
View all books & articlesClean Architecture
Y'all should read this book even if you do not like his tweets.
The Age of Surveillance Capitalism
A lot of the points now might feel like public knowledge, but this book was one of the first to systematically analyze how big tech companies manipulate personal data for profit. For better or worse this information is going to shape your thoughts on privacy and the digital economy.