I build the decision systems that businesses trust.
The models change every year. Great decision systems endure.
Vision
AI is moving from something you operate to something that operates for you. The next decade belongs to organisations whose systems can anticipate, simulate a decision before making it, and improve themselves in production. Advantage will not come from better models. It will come from the architecture around them: the decisions they are allowed to make, the controls around those decisions, and the trust they earn in the real world.
Expertise
Decision Architecture
Production & Agentic AI
AI Governance & Responsible AI
Voice & Multimodal AI
Building AI Organisations
Applied across lending, insurance, wealth, retail and pharma.
Research Interests
- Reward architecture: learning loops that turn live production signal into continuously improving models.
- World models and simulation-first decisioning: agents that rehearse a decision before they act.
- Self-reflective agents that learn between episodes, not only during them.
- Speech stacks for Indian-context conversational AI: prosody, turn-taking, code-mixed language.
- Evaluation built from production traffic rather than public benchmarks.