Employee Spotlight: Darsh Selarka

Meet Darsh, QuadSci's founding Forward Deployed Engineer, who is helping customers turn complex data into actionable intelligence. With more than four years of enterprise AI deployment experience at C3 AI, he's focused on making QuadSci implementations faster, simpler and more scalable, while ensuring customers can put Growth AI insights to work where they already operate.

Why QuadSci?

I believe in what QuadSci is doing and how it's doing it. The platform learns from how customers actually use a product and turns that into early, explainable predictions of churn and expansion, giving teams a real head start on their most important relationships. I believe just as strongly in the team building it. Our founders bring deep experience and knowledge of this space, along with real conviction in solving it, and everyone I met on the team was excited to build something great.

What's your background?

I studied Computer Science at Indiana University Bloomington. I then spent more than four years as a Forward Deployed Engineer at C3 AI, leading enterprise AI deployments end to end for customers in pharma, oil and gas, retail, energy and more. Every industry brought different data, different stakeholders and a different definition of success. Across all of them, my job was the same: get the solution live in production and make sure the people using it actually got value from it.

What are you building right now?

As QuadSci's founding Forward Deployed Engineer, I'm helping build the FDE function from the ground up with our CTO, Maggie. That means defining how we take a customer from first data handoff to live predictions, and turning each deployment into something faster and more repeatable than the last.

I'm also championing adoption of the QuadSci MCP Server across our customers. It lets teams ask questions of their QuadSci insights directly from AI assistants like Claude, instead of logging into one more dashboard. The insight that changes a decision is the one that shows up where people already work. Getting MCP into customers' hands drives adoption for us and saves time for them.

What does accurate and useful intelligence look like to you?

One thing deploying AI has taught me is that a model can be accurate and still go unused. If the people receiving a prediction can't see why it flagged an account, they won't act on it, however good the numbers look. To me, useful intelligence is something a revenue team understands at a glance and can act on that same week. A score on its own isn't enough. It needs the reason behind it and enough lead time to actually change the outcome.

One thing people should know about you

I'm a curious learner at heart, and I try to live by KISS: keep it simple. If I can't explain something in plain words, I probably don't understand it well enough yet. Lately I've been getting into chess, which has sharpened how I think about pattern recognition: reading the shape of a position before calculating every move. Outside work, I've been spending a lot of time on the tennis court lately and have been loving it.