Built by people who have shipped real systems in regulated, high-stakes verticals.
Kalman AI is small by design. The work is done by people who have built and operated AI systems in real environments — not just prototyped them.
Dr Jehill Parikh
Physicist · Applied neuroscience · World models since 2015
Trained as a physicist, with PhD-level NeuroImaging research and a long collaboration with neuroscientists translating brain research into tools clinicians and industries actually use.
Built and shipped generative and decision-under-uncertainty systems for clinical and industrial problems years before transformers and modern LLMs existed. The conviction behind our engineering bets is not opportunistic.
Now turning that decade of applied work into Kalman AI: a consulting practice that takes AI from pilot to production for industries that can’t afford to get it wrong.
Three things make Kalman AI different from a typical AI consultancy.
A founder who has shipped, not just spoken
Production AI agents in regulated verticals built from zero, by the same hands that will design your engagement. A decade of work on world models and decision-under-uncertainty systems before modern LLMs existed.
An engineering-first practice
We don’t sell platforms or wrappers. We design, build, deploy, and maintain agents — and stay accountable for outcomes after launch. Operations is part of the engagement, not a hand-off.
A senior team — small by design
Every engagement is led by people with production experience in the relevant vertical. The first conversation you have is with the founder, not a sales team. We say no quickly when the fit isn’t right.
Based in India — with mandates across the country and partnerships further afield.
India is conspicuously underbuilt in autonomous-decision AI for regulated verticals — independent CTO data suggests only around 6% of Indian CTOs are actively working in this space, even as global demand for exactly that capability is rising fast. We are deliberately positioned in the gap between that scarcity and that demand.