Abhinav Soni is Co-Founder of BentoLabs AI (YC P26), building the monitoring and self-learning infrastructure layer for production AI agents, helping teams detect silent failures, debug regressions, and continuously improve agent performance.

Abhinav studied at ESCP Business School and gained early-career experience at INDmoney and UnBook before becoming the first hire at Emergent (YC S24), where he worked on agentic systems at scale.

As hire #1 at Emergent, Abhinav helped the company hit the top spot on SWE-Bench and scale from $0 to $100M ARR in just 8 months, one of the fastest growth trajectories in AI startup history.

Alongside co-founder Kaushik, Abhinav built and operated production coding agents at Emergent that served over 5 million users, giving the BentoLabs team deep first-hand insight into the hardest problems in agent reliability and observability.

BentoLabs' self-learning engine has demonstrated a 2.6x lift on ARC-AGI-3 and improved Terminal-Bench 2.0 pass rates from 42.2% to 52.4% using the same model, tools, and budget, signaling strong early technical differentiation in the AI infrastructure space.