paraskapoor.com

Founding Engineer · Production AI

Paras Kapoor — Founding Engineer · 0→1 Product Architect

I take products from zero to production.

Founding engineer on the official Atomic Habits app. Now I build autonomous AI agents and RAG systems that ship to real users — not demos.

currently — architecting AI agent systems
Atomic Habits app/Telkomsel Indonesia/Hay House/−80% latency/RAG · Agents · LLMs/0→1 founding engineer/Atomic Habits app/Telkomsel Indonesia/Hay House/−80% latency/RAG · Agents · LLMs/0→1 founding engineer/

01Selected Work

Things I took from idea to production.

02Philosophy

Production is the only truth.

01

Boring beats clever.

Most teams reach for microservices years too early. A clean monolith beats a distributed mess for almost every product that isn't yet at scale.

02

Most AI is demo-ware.

The model is the easy part. Retrieval, evals, edge cases, and cost are the job. I ship AI that survives contact with production.

03

Cost is engineering.

An architect who can't tell you what it'll cost isn't finished thinking. Estimation and cost analysis are part of the design, not an afterthought.

04

Prod is the only truth.

You don't understand a system until you've run it in production and watched it break. Observability before scale; idempotent and fault-tolerant by default.

03About

I build products that survive production.

Software engineering isn't just writing code. It's making thousands of technical decisions that determine whether a product remains reliable, maintainable, and scalable long after it's shipped.

I enjoy working where engineering meets product — owning it end to end. From backend systems and cloud infrastructure to AI and the frontend people actually use, I help turn ideas into production software that solves real problems.

I believe technology should create business value, not unnecessary complexity. Every architectural decision is a trade-off between reliability, performance, cost, and developer experience.

My recent work focuses on backend platforms, cloud-native systems, AI applications, and building software that teams can confidently operate, evolve, and scale.

Explore how I think

04Writing

Notes from production.

Why most AI products die in production

AI Engineering

soon

Cost is an engineering decision

Architecture

soon

You don't need microservices yet

Systems

soon

05 — Contact

Building something that has to work?

I'm not job-hunting — but I'll always make time for a hard problem, an early-stage product, or a conversation worth having.

hello@paraskapoor.com