Mallik Cheripally

Staff Engineer at Deepen AI, Hyderabad

I build complex products that stay clear, fast, and reliable as they grow.

I take products from a blank page to production, and own the product decisions, the architecture and the performance along the way.

Open source

Four packages I maintain in the open, covering local coding-session management, color, accessible focus behavior and JavaScript utilities.

All open source
  1. A local session manager for Codex and Claude Code. Review, understand, and clean up sessions safely, without anything leaving the machine.

    • Local
    • Browser + CLI + MCP
    • AI
    • Session Manager
    npm install --global session-steward
  2. A zero-dependency color toolkit for conversion, manipulation, harmony generation, parsing, and contrast analysis.

    • Color science
    • Zero dependencies
    • Accessibility
    npm install colore-js
  3. Focus-management primitives for React: trapping, restoration, keyboard order, and announcements.

    • Accessibility
    • Keyboard interaction
    • React
    npm install react-refocus
  4. A fast, focused deep-equality utility for JavaScript and TypeScript.

    • Deep equality
    • Type coverage
    npm install deep-equal-js

About

I like the part of the job where the answer isn't written down yet.

I'm Mallik Cheripally, a Staff Software Engineer at Deepen AI, based in Hyderabad. For more than a decade I've built complex web products, product platforms, and developer tools, from shaping the first version to keeping mature systems clear, fast, and dependable as their data, teams, and requirements grow.

Often the starting point is an idea, not a specification, so someone has to decide what the product is before anyone can build it. I do both parts. The hardest call is usually what not to build yet, and the states that matter most are the difficult ones: the empty ones, the failing ones, the ones with more data than expected. I build it myself, so if a decision is wrong I find out while it's still cheap to change, and I stay on it while real data and usage find the constraints that weren't visible at the start.

I joined Deepen in 2018 as a senior frontend engineer, took more than five of its products from an empty repository to production, annotation and calibration among them, and was promoted to Staff in 2022. These days I own products end to end, from AI and data pipelines to the interface, guide the engineers I work with, and set the practices we build by. Before Deepen, I did product engineering at Knex and Java applications at TCS.