From Startup Founder to Healthcare AI Engineer: My Journey

Eighteen months ago, I was sitting in a small coworking space, staring at a dashboard that showed ReachGig's active users. I was the Founder and CTO, and I had built this freelancing platform from the ground up. It was my baby. Every line of code, every marketing campaign, every user interaction—it all had my fingerprints on it.

Today, I'm a Founding Engineer at Mantys Healthcare AI, a Y Combinator company. Instead of worrying about user acquisition costs for freelancers, I'm building AI systems that process medical procedures for hospitals.

It’s a shift that might look surprising on paper. Why go from being your own boss to an engineer at someone else’s startup? But the truth is, this transition has been one of the most rewarding chapters of my life. It taught me that titles matter less than the problems you’re solving, and that the skills you build as a founder are your secret weapon, no matter where you go next.

This is the story of that pivot.

The ReachGig Chapter: Building in the Trenches

ReachGig was my crash course in... everything.

When you're a technical founder, you don't just write code. You architect the entire system, sure—I used Next.js and AWS Serverless because I needed speed and low costs—but you also become the product manager, the customer support agent, and the head of sales.

I remember nights spent debugging a race condition in our notification system (we finally hit a 99.5% delivery rate, which felt like a massive victory at the time) and then waking up four hours later to pitch to a potential partner. It was exhausting, but exhilarating.

We had wins. We reduced our infrastructure costs by 40% because we had to—we didn't have endless VC money to burn. We built a React Native app that users actually liked. We got traction. But slowly, a hard reality set in.

The Pivot Decision

Leaving your own company feels a bit like giving up on a dream. But as I looked at the market, I realized that the freelancing space was becoming a war of attrition. We were fighting giants like Upwork and Fiverr, who had deeper pockets and massive network effects. To win, we needed capital I wasn't sure we could raise, or a differentiator that we hadn't quite found yet.

At the same time, I was feeling a different kind of itch.

After two years of "broad" problem solving—marketing, ops, legal—I missed "deep" technical work. I wanted to sink my teeth into a problem that wasn't just about moving data from A to B, but about solving something hard.

Healthcare AI was exploding. It was complex, heavily regulated, and high-stakes. It wasn't just about convenience; it was about patient care. When I met the team at Mantys—smart, driven, and backed by YCombinator—it clicked. This was the room I wanted to be in.

Culture Shock: New Domain, New Rules

Changing tracks to Healthcare AI wasn't just a new job; it was a new world.

In the startup world, "move fast and break things" is the mantra. In healthcare, if you break things, people get hurt (or you get sued).

My first three months were a humbling experience. I wasn't just learning a new codebase; I was learning a new language. HIPAA compliance, ICD codes, CPT procedures, prior authorization workflows—it was a firehose of information.

I found myself reading FDA regulations instead of Hacker News. I spent hours with our domain experts, asking "stupid" questions about how insurance claims actually work.

But here's the cool part: the engineering fundamentals didn't change.

The Founder Advantage

I realized quickly that my time as a founder gave me a unique edge as an engineer.

I saw the whole board. Most engineers optimize for code quality or performance. As a former founder, I instinctively optimized for value. When we built our Prior Authorization system, I didn't just ask "how do we build this model?" I asked "how does this fit into the hospital's existing workflow? How does this save them money?"

I wasn't afraid of the "boring" stuff. Founders know that the shiny AI model is only 10% of the solution. The other 90% is data cleaning, error handling, compliance logging, and integration. I dove into those areas because I knew that's where the product actually lives or dies.

I knew how to scrappy. At Mantys, we needed to evaluate our LLMs, but we didn't have months to build a research-grade framework. So I built one using what we had—Cove and Gevals. We got to 98% accuracy not by being perfect, but by being iterative.

What I'm Building Now

Today, I'm doing the deepest technical work of my career, and I love it.

We're building systems that automate prior authorizations for over 120,000 procedures. We've built financial reconciliation tools that have recovered millions of dollars for providers. It’s high-impact work.

We achieved 98% accuracy in our data extraction, which sounds dry until you realize that every percentage point represents a human being who doesn't have to manually data-enter a complex medical form.

To Founders Looking for What's Next

If you're a founder thinking about transitioning back to an engineering role, or joining another startup, here is my advice: Do it.

Don't view it as a step back. It's a step sideways into a new arena where you are deadlier than you think. Your ability to think like a business owner while coding like an engineer is a superpower.

Look for problems that scare you a little. Look for teams that are smarter than you. And don't worry about the "Founder" title. The best title you can have is "The person who gets important shit done."

Healthcare AI is just getting started. We are barely scratching the surface of what's possible when you combine advanced LLMs with medical expertise. It's a messy, complicated, difficult field.

And I wouldn't want to be anywhere else.

---

Harshavardhan is a Founding Engineer at Mantys Healthcare AI. Previously, he was Founder and CTO of ReachGig. If you're navigating a career pivot or interested in the wild world of Healthcare AI, hit me up on LinkedIn/Twitter.