About: Mohith G

The short version

I'm a Product Engineer in the way that label means something today: a generalist who knows enough at every layer of the stack to make the call about what to build, not just how to build it. AI shows up in two ways in my work: as a tool I use to ship faster (agents in the loop), and as a layer in the products I ship (translating real systems into something users can act on). I've been doing this professionally since 2014. Most of that time has been at venture-backed startups, several of them Y Combinator companies, all of them either growing fast or trying to.

Today I'm the first engineer at PortfolioPilot, an SEC-registered AI financial advisor. We built a hedge-fund-grade quantitative engine ground-up: macro analysis, multi-model risk and return forecasting, the kind of analysis a private wealth manager runs. The AI sits on top, translating that math into language a normal investor can act on. Before this, I was a founding engineer at IntuitionAI (which Domino Data Lab acquired), and I spent three years on Domino's core platform. Before that I spent over a year building early product surfaces for Rippling, back when "production React" was a sentence that needed a footnote.

How I work

I have strong opinions about how to ship software. They're not unusual opinions; they're just not commonly applied with discipline. They come down to:

Ship the smallest thing that proves the next decision.

Most engineering effort wasted at startups is wasted on building too much before the next signal arrives. The hardest discipline in early-stage work is figuring out what would change your mind, and building only the thing that surfaces that signal. Then iterate.

Build the substance, then the surface.

The most common AI-product failure I see is teams shipping the LLM wrapper before they have anything underneath worth wrapping. The hard part of an AI product is almost never the prompt; it's the system the prompt is summarizing. At PortfolioPilot the substance is a hedge-fund-grade quantitative engine, and the AI is the layer that makes it legible. I wrote about this in the PortfolioPilot case study.

Restraint is craft.

Frontend craft is not tricks. It's restraint. Backend craft is not patterns. It's restraint. The best code I've written and the best code I've reviewed is the code that didn't add the abstraction yet. I learned this the hard way, in a 2016 React codebase that should have shipped in half the time it did. Story.

The boundary between "design" and "engineering" is mostly imaginary.

I write design docs. I draw architecture diagrams. I prototype UI. I do code review. I write the prompts. I run the SQL. Treat the boundaries between these as friction to remove, not roles to defend.

The credentials

Computer Science gold medalist (batch topper) from Saveetha Engineering College (Anna University). Eleven years professional experience across AthenaHealth, Codebrahma, Rippling, Navya, IntuitionAI, Domino Data Lab, and Global Predictions. Open-source contributions across the npm ecosystem. The full timeline is on the home page, drawn as a directed acyclic graph because I think in DAGs.

Off-hours

I live in Bengaluru. I read more books than I should. I'm into the kind of technical writing where someone shows you how a system actually works, not how it's marketed (Julia Evans, Dan Luu, Kent C. Dodds, Lilian Weng). Off-hours, I tinker with Raspberry Pi: home automation, sensor rigs, the occasional embedded oddity. There's something grounding about a system that fits in your hand after a workweek spent in distributed cloud.

If you're hiring

I'm open to founding-engineer or first-AI-engineer roles at AI-forward seed-to-Series-A teams. I'm also happy to talk shop about agent stacks, building real systems beneath the LLM, or how to think about taking an AI feature from "it sometimes works" to "we'd ship this."

The fastest way to reach me is email. The most efficient way is to chat with my career. There's a button somewhere on this site for that.