Ashish spends his days where ideas become reality, partnering with the team, then the founders to transform early concepts into products people can actually use, and want to buy. Leveraging AI to accelerate development and experimentation, he helps ventures move from idea to MVP faster, without losing sight of the customer problem they're solving. Whether he's designing solutions, building new features, or testing the latest iteration, no two days look the same. Having helped launch Tracksuit and every New+Improved venture since its earliest stages, he's developed a deep appreciation for the messy, fast-moving nature of building something new and the constant learning that comes with it.
What sparked your interest in building software products?
I didn't grow up knowing I wanted to work in tech. What I did know was that I enjoyed logic and problem-solving, and programming turned out to be a natural fit. I loved the satisfaction of building something that solved a real problem, and over time that curiosity gradually became a passion for building products.
From when you built Tracksuit’s MVP to today, how has AI changed the way you build products?
AI has fundamentally changed how I build products. Instead of spending most of my time writing code from scratch, I now focus more on product outcomes, architecture, and tests. AI accelerates implementation significantly, so I spend much more time thinking about testing and ensuring the product behaves as intended. The biggest challenge is making sure AI is building exactly what I want and avoiding becoming overly reliant on generated code that isn’t solving the true problem we are working on.
How do you stay up to date with how quickly technology is changing?
YouTube is one of my main sources for keeping up with new technologies. The tech community moves quickly, and creators often publish useful content within days of major announcements. A few I recommend are Better Stack and NetworkChuck. I also learn by experimenting with new tools and building small prototypes, which helps me understand what's genuinely useful.
What’s the most valuable thing you’ve learned from building multiple ventures?
There is never a perfect technology stack or a perfect decision.
Every venture teaches you something, and there are always things you would do differently next time. The exciting part is that each new venture is a fresh opportunity to apply those lessons.
Handing products over to dedicated teams when the ventures officially launch and hire their teams is always both rewarding and a little bittersweet.
As AI speeds up development, what can’t it replace and where do bottlenecks move?
AI can't fully replace human judgement, creativity, or intuition, which we lean on heavily in how and what we build!
People bring opinions, context, and empathy to our products, which shape the user experience. Those human perspectives are what make products unique, and I think that will remain difficult for AI to replicate.
Product clarity, validation, and quality assurance become really important. Actually far more important than previously. The key challenge is ensuring that AI is producing exactly what we intend it to, that the final product genuinely solves customer problems and that we haven’t got carried away with AI’s lack of judgement.
How do you balance speed with quality when building early-stage products?
In early-stage products, speed is essential because the goal is to build, learn and iterate quickly.
I spend less time perfecting code and more time investing in testing and validation. AI helps accelerate development, but strong testing and rapid feedback loops ensure quality doesn't suffer. The goal is to build the right product quickly rather than a perfect product slowly.
What's one AI tool you couldn't imagine working without now? And why?
AI coding assistants such as Cursor and Claude have become indispensable to my workflow. They dramatically speed up development, help with debugging, and allow me to prototype ideas much faster.
What I am most excited about and more importantly, they free up time to focus on product thinking and user experience rather than repetitive implementation tasks. I think leveraging the tools' capabilities in the right way is more important. For example, using MCPs everywhere to connect the system. The cool thing we did was we spun up our own DB MCP so Claude/Cursor can debug the errors very easily.
What's a technical shortcut that's almost never worth taking?
Ignoring security or best practices for the sake of speed. Tech debt accumulates quickly and usually becomes expensive later.
But, a shortcut that is almost always worth taking is using AI to accelerate the first version of a product, while still maintaining strong testing and engineering fundamentals. We can get the basics done fast, then be able to finesse the features that need human judgement.
What does a day in your role look like?
No two days are the same, which is part of what I enjoy most. A typical day involves working with founders and early teams to understand requirements, designing solutions, building features, testing AI-generated code, and testing new ideas.
What's the most challenging technical problem you've had to solve and how did you go about it?
One of the biggest challenges in early-stage products is balancing speed, scalability, and uncertainty at the same time. Requirements change quickly, so it's important not to over-engineer solutions too early. I usually break problems into smaller pieces, validate assumptions early, and build systems that can evolve as the product grows.
What keeps you excited about software development when technology changes so quickly?
The constant change is actually what keeps it exciting. There is always a new tool, framework, or idea to explore. I enjoy the combination of creativity and problem-solving, and I still find it incredibly rewarding to turn ideas into products that people can use.
What’s the best piece of advice you’ve ever received?
"Spend money on experiences rather than things."
We hope you enjoyed hearing from Ashish! Connect with Ashish here!
