Meet Sunil Pai

Software Developer

Sunil Pai is a software developer and technical writer focused on the modern JavaScript and web platform. His work centers on stateful serverless infrastructure, Cloudflare Workers and Durable Objects, realtime systems, local-first architecture, React, and AI agents, with writing that explores how these primitives can make complex applications simpler to build.

Sunil Pai (digital)
Sunil Pai
sunilpai.dev
June 14, 2026

Summary of Sunil Pai’s portfolio site

Sunil Pai’s website is a minimalist personal portfolio/blog titled “Solving the decision problem.” The About page is intentionally sparse — “Hi, I’m Sunil Pai.” — so the clearest picture of him comes from his writing.

Sunil is a software developer and technical writer focused on the modern JavaScript/web platform: React, bundlers, serverless infrastructure, Cloudflare Workers, Durable Objects, realtime systems, local-first apps, and AI agents. His portfolio is less a traditional projects page and more a chronological record of technical essays, experiments, architectural arguments, and product ideas.

What he works on

A major through-line in Sunil’s recent work is stateful serverless computing. In “The next evolution of serverless is stateful”, he explains his work on PartyKit, a realtime/stateful serverless platform powered by Cloudflare Durable Objects. He frames PartyKit as a way to make serverless feel more like simple single-machine programming, while still scaling globally. Use cases he highlights include collaborative editing, multiplayer apps, workflows, chatbots, sync engines, local-first apps, and games.

He continues this line of thinking in “Durable Objects are Computers / Containers / Servers”, where he describes Durable Objects as cloud “computers” or actors: addressable units of compute with state, storage, WebSockets, and long-lived behavior.

His most recent technical writing has shifted heavily toward AI agents as long-running, stateful software. In “Full Stack AI Agents”, he argues that AI agents should not just have input/output/state, but also their own UI or admin surface — essentially making each agent a tiny full-stack app. In “AI agents are local first clients”, he connects agents to sync-engine/local-first architecture, suggesting agents should sync state and perform actions much like browser or mobile clients. In “an event bus for ai agents”, he explores agents reacting to queues/events and deciding what to do next.

He also writes extensively about Cloudflare Workers as infrastructure for AI products, especially Durable Objects, Workers AI, AI Gateway, D1, R2, Vectorize, and browser automation. See “A list of reasons why you should be using Cloudflare Workers for building your AI agent infrastructure/product/personal assistant”.

Web platform and JavaScript interests

Sunil has a long-standing interest in the JavaScript ecosystem and frontend architecture. Older posts cover topics like:

His writing style is informal, opinionated, and architecture-heavy: he tends to reason from first principles, sketch APIs, and explain why certain primitives make software simpler.

Current themes

Across the site, Sunil seems especially interested in:

  • Stateful serverless / actor-model infrastructure
  • Cloudflare Workers and Durable Objects
  • Realtime and multiplayer applications
  • Local-first/sync-engine architecture
  • AI agents as persistent, inspectable, full-stack systems
  • React rendering, streaming, Suspense, and performance
  • Developer experience, devrel, and how engineers learn

A recent essay, “developer relations after the cheat code machine”, reflects on how AI coding tools change developer education. His argument is that API knowledge is becoming cheaper, while judgment, taste, and “watch me work” apprenticeship-style learning are becoming more valuable.

Social links

The homepage links to several profiles:

Overall, Sunil Pai’s portfolio presents him as a deeply technical web-platform engineer and builder, with recent work centered on PartyKit-style stateful serverless systems, Cloudflare infrastructure, and the architecture of AI agents.