Utkrusht AI raises seed round for 'watch-them-work' hiring platform
What's the deal? Vadodara-based startup Utkrusht AI has raised a seed funding round to scale what it calls the world's first "watch-them-work" assessment platform for technical hiring. The platform drops candidates into live production environments — a broken payment microservice, a failing Kubernetes cluster, a slow API — and records how they solve real problems in 30–45 minute sessions.
Rather than relying on coding tests, take-home assignments, or AI interview tools, Utkrusht lets hiring managers observe candidates executing actual on-the-job tasks. Evaluations cover technical execution, problem-solving approach, decision-making, and — crucially — how candidates use AI tools during the session.
Why now? AI has fundamentally changed what engineering work looks like. Traditional assessments either ban AI use or can't detect it, making it harder to distinguish between candidates who genuinely understand a problem and those who can't explain the code they generated.
Utkrusht tracks exactly where, how often, and how effectively each candidate uses AI — turning a blind spot in modern hiring into a signal. Tech hiring still takes four to eight weeks on average to fill a role, and companies report the platform cuts that by roughly 70%, with hires typically closing in under 14 days.
What could go wrong? The platform's value hinges on the quality and realism of its simulated environments. Building and maintaining a "massive infrastructure" of live production tasks across backend, DevOps, full-stack, AI engineering, and data engineering is operationally complex and expensive — especially at scale.
Utkrusht also faces a crowded market. Established players like HackerRank and Codility have brand recognition and enterprise contracts. Convincing hiring managers to abandon familiar tools for a fundamentally different workflow is a steep sell, even with strong data behind it.
The signal: The rise of AI-assisted coding has broken the assumptions underlying most technical assessments. If you can't tell whether a candidate wrote the code or just prompted it, test scores become unreliable. Utkrusht is betting that observing process matters more than grading output — a shift from "did they get the right answer?" to "how did they think through the problem?"
The company says it has assessed more than 10,000 candidates, with backtesting showing that its shortlisted candidates consistently turn out to be the strongest hires. If that holds at scale, the approach could reshape how companies evaluate technical talent in an AI-augmented world.
Read more: techcircle.in
