About

I'm Arbaz. For ten years I have been building distributed systems at scale. Live streaming, real-time audio, social graphs, payments and ticketing, and crypto exchange infrastructure. As a founding engineer at Turnip I scaled the platform to 25M users.

Since 2024 the work has moved to AI models and what surrounds them. GPU inference, fine-tuning, retrieval, and autonomous pipelines. I learn it by shipping it, so this year I released a language model, built a media channel that runs itself, and put apps on the App Store.

Work

Turnip

Principal Software Engineer, January 2021 to present

One startup, four pivots: mobile-game streaming, Clubhouse-style live audio, a creator social network, and now Zaps, an AI design suite.

  • Built and ran the backend behind 25M users. 20+ services on Kubernetes, written in Node.js and Go, fronted by Kong with Kafka and Redis carrying traffic.
  • Wrote the game-streaming stack, RTMP and FFmpeg ingest running 50K streams a day at sub 5s latency.
  • Implemented live audio rooms, where MQTT and Firebase kept every listener's state in sync for 2M people a month across 70M minutes of calls.
  • Built a social network with a news feed, discovery, DMs and connections over Neo4j for second-degree retrieval, indexed in real time by Elasticsearch on an event-driven Kafka core.
  • Shipped AI photo editing, face swap, retouch and background tools on ComfyUI, orchestrated with LangGraph and n8n on Kubernetes.
  • Cut GPU spend by 80% by moving generation onto a distributed GPU network while still returning images at sub 2s, and another 70% off infrastructure by moving to serverless.
  • Made the template library searchable by meaning rather than by tag, semantic search over embeddings retrieved by vector similarity.
  • Set up an engineering platform where ten product people shipped 17 internal tools without a single engineer commit.

MouseBelt

Senior Backend Engineer, April 2020 to January 2021

A centralized cryptocurrency exchange, and the crypto work around it.

  • Built the exchange backend. Accounts, KYC, sessions, markets, order flow, fee schedules, referral commissions, API keys with IP allowlists, and failover.
  • Worked on the low-latency core, a matching engine written in C with its own gRPC and WebSocket access layers, market price feeds, and trade history services.
  • Multi-chain custody. Deposits and withdrawals across Bitcoin, Ethereum, EOS, Cosmos, ICON, and Binance Chain.
  • Built a multi-chain DEX aggregator with price scrapers across BSC, ICON, and Moonbeam, plus a blockchain indexer for on-chain history.

Paytm Insider

Lead Backend Engineer, 2017 to April 2020

  • Led the team behind movie ticketing that held 500+ requests a second.
  • Designed the logistics architecture. Delivery opt-in, order tracking, and the internal tools for managing shipments.
  • Built the full e-commerce flow. Cart, payments, inventory, and taxation.
  • Ran it all as AWS microservices on Kubernetes, with serverless functions on Lambda and Kinesis carrying the on-demand work.
  • Led a team of 7 and interviewed 100+ engineers for backend and DevOps roles.

Ithaka

Intern, 2016 to 2017

A travel-planning startup. Built its omni-channel chat server bridging the app, Facebook Messenger, and support agents over Socket.IO and Firebase, with conversation handoff and Elasticsearch-backed search.

Provakil

Intern, 2016

Legal-tech. Wrote scrapers and fetchers that pulled daily cause lists from Indian courts, including the Delhi High Court and the national consumer courts, into structured case data.

Now

Right now I build as an indie dev. Idea, product, distribution, growth. Day to day the work is automation pipelines, pushing language, diffusion and retrieval models to see how cheaply they can run unattended before the quality gives out.