About — Depend.AI

I started Depend.AI after twenty years building AI systems in healthcare: at the research bench, in the engineering function, then at the executive table. What I see now in conversations with funds, and their portfolios: boards being handed AI quotes and roadmaps by agencies, and no one in the room they trust to tell good advice from bad.

Depend.AI is the firm I built for those rooms.

Saurabh Johri, PhD Founder & Principal, Depend.AI

Background

Advisory grounded in the work, not the slides

Two decades of building AI inside healthcare.

01
Founder & Principal — Depend.AI Investors and operators · Regulated industries Principal-led AI advisory and engineering for investors and operators in regulated industries: AI due diligence for funds, post-investment AI strategy for portfolios, and architecture-led engineering for high-stakes deployments. Active engagements in UK home insurance and consumer mental health technology.
02
Chief Scientific Officer — Babylon Health Healthcare AI · NHS scale · Seed → IPO → public markets Architected and led the deployment of clinical AI systems used by payors, providers, and patients at scale, under NHS contracts, MDR oversight, and the scrutiny of regulators, clinicians, and institutional investors simultaneously. Babylon went from growth-stage through IPO and into public markets during this time. Applied AI leadership at every stage of the company lifecycle, and the experience of building AI where failure has real human consequences.
03
AI Research Laboratory Leadership Peer-reviewed research · Patents in medical AI Led AI research laboratories. Peer-reviewed publications in applied AI. Patents in medical AI: counterfactual measures for clinical diagnosis, probabilistic graphical models with causal structure, and conversational AI architecture.
04
PhD — Imperial College London Population Genetics & Computational Biology PhD in Population Genetics and Computational Biology (2003–2007). Research grounding in probabilistic models, biological systems, and computational methods. The foundations that underpin modern AI.

Why Depend.AI

Where most AI advisory falls short

Most organisations in regulated markets hit a recurring problem when they look for AI advisory. They find strategy firms producing frameworks and governance decks without the engineering depth to know whether the advice can be built to the standard a regulator sets. The alternative is engineering shops moving fast on prototypes and integrations, without the accountability that a board or regulator demands when those systems are running in production.

Neither half is sufficient when the AI system has to withstand that kind of scrutiny.

That is the gap Depend.AI exists to close. The advisory carries weight because it draws on real production engineering. The engineering is accountable because it is shaped by the strategy that regulated markets require.

About the firm.

Advisory grounded in the work, not the slides

Where most AI advisory falls short

Build AI your organisation can depend on.