What if drug design didn’t depend on pre-existing biological data?

In this episode of Mendelspod, Theral sits down with Aridni Shah, co-founder and CEO of Immunito AI, a Bengaluru-based biotech startup reinventing how we develop antibody therapies. Shah and her team are using artificial intelligence to design antibodies from scratch, bypassing traditional animal-based and data-heavy methods. “We no longer need to rely on pre-existing data,” she explains. “We’ve gone to the fundamental atomic level and learned basic physics, chemistry, and maths—why do two proteins interact?”

* 0:00 Discovering new antibodies without relying on biological processes

* 4:45 The platform - a structure first approach

* 10:00 The business model

* 15:05 Bringing mathematics and physics to biology

At the heart of Immunito’s innovation is a proprietary mathematical transformation that translates raw 3D structural data into rich, generative models capable of designing antibodies for even “undruggable” targets. It’s an ambitious approach to rational drug design that prioritizes data efficiency, leveraging physics-based complementarity rather than brute-force sequence matching.

Still early in development, Immunito is already seeing promising lab results and aims to partner with pharma while building its own pipeline. “We should have signed a few multimillion-dollar multi-target deals,” Shah says of their goals over the next few years, “and be looking to progress towards IND.” This conversation opens a compelling window into how AI, when grounded in mathematical and physical intuition, may radically transform biology.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.mendelspod.com/subscribe