I'm a backend engineer with 6+ years of experience in building software, currently one of the longest-serving engineers on the tech team at Seez — four years in, with end-to-end knowledge of the systems that power the company. I own the backend behind Seezar, an AI chatbot platform live across 100+ dealership groups in Europe, the Americas, Australia, and New Zealand.

Seezar (and its internal sibling, Seezar-Assist) processes 15M+ messages through a WebSocket architecture I built from the ground up, and was the primary driver behind $3M+ in revenue and Seez's 2025 acquisition by Pinewood AI.

const arbaz = {
  role: "Lead Fullstack Developer",
  focus: ["AI chat systems", "distributed backends", "developer velocity"],
  stack: {
    core: ["Node.js", "PostgreSQL", "GraphQL"],
    cloudServices: ["AWS Lambda", "EventBridge", "SQS", "S3"],
    exploring: ["Python/FastAPI", "LLM tooling", "RAG pipelines","Go"]
  },
  currentlyLearning: "how far I can push AI-native backend architecture"
};

• Real-time AI chat infrastructure — Designed and scaled the WebSocket layer connecting frontend clients to an LLM-powered chatbot, handling millions of concurrent conversations across a multi-tenant, multi-region deployment.
• Multi-channel messaging system — Built WhatsApp, SMS, and email notification pipelines using Infobip.
• Human-in-the-loop AI handover — Architected a Redis pub/sub fan-out system enabling seamless AI-to-human handoff across Dockerized instances, so dealers can step into a conversation the moment it matters.
• In-house analytics engine — Took over and scaled a high-throughput analytics module (SQS-backed event ingestion) that replaced a third-party analytics vendor, keeping company data fully in-house.
• Core platform API — Contribute to a GraphQL-based, serverless (Lambda) modular monolith powering dealership, inquiry, and user-facing services at Seez.

I like understanding systems from the inside out — not just using tools, but knowing why they work the way they do.

• Database internals: B+ Trees, LSM Trees, Bloom Filters, WALs, replication & sharding
• LLM integration patterns: RAG pipelines, async orchestration, agent architectures
• Sharpening distributed systems fundamentals for senior/staff-level backend design

• Feature-based folder structures over rigid MVC layering
• Conventional Commits, always
• Believer in shipping systems that are boring in the best way: predictable, observable, and easy to hand off

Always happy to talk backend architecture, AI infra, or the occasional swing trade 📈