I am an LLM Psychologist and Head of Innovation at Straive – I study how AI systems behave, fail, and can be verified.
I research how AI behaves
I publish reproducible experiments on model behaviour, prompting, hallucinations, evaluation and coding agents.
… And turn the findings into working prototypes
At Straive I lead rapid AI R&D: testing new capabilities, building prototypes, and turning what works into reusable tools and playbooks.
I am faculty at IIT Madras, teaching Tools in Data Science in the IITM BS programme.
I redesigned the course for an AI-available world…
Prompts instead of static content, “Ask AI” assessments, open-ended problems and automated evaluation – teaching orchestration, debugging, verification and judgment.
… And I grade with LLMs, not just people
In one experiment, LLM-assisted grading cut the evaluation of 500 code submissions from about two weeks to about two hours – matching my judgment about as closely as TAs did.
My public work is an open research lab – open code, experiments, data stories and talks.
I publish on LLM psychology and reliability…
I write on model behaviour, coding agents, analysis and AI-native education, with reproducible code and data behind each result.
… And speak to practitioners and policymakers
Recent talks and workshops include TEDx, PyCon, FOSSASIA, DataHack Summit, IIT Madras, Harvard Graduate School of Education and LBSNAA.
Earlier, I co-founded Gramener and was its CEO / Chief Data Scientist until Straive acquired it in 2023.
Before that, consulting and engineering
Principal Consultant at Infosys Consulting, London; consultant at The Boston Consulting Group, Mumbai; and software engineer at IBM, Bangalore.
