The Talent Question — Summary

Kavitha graduated from IIT Madras with research in small-model optimization, two offers, and a rational choice. San Francisco paid in dollars, with stock options that would vest into an amount she had never held. Bangalore offered a mission to build Indian-language models for public health and a salary in rupees. She took San Francisco. She was not wrong to.

The talent question is not whether people like Kavitha exist. India produces them abundantly. The question is where they go, and whether the places that need them can make staying rational. The frontier recruits the best because the compensation gap is wide, the talent is rare, and the window is short.

What the floor needs is different talent: people who can deploy, fine-tune, maintain, and govern models in specific contexts. Not less skilled but differently skilled, and undervalued because the work is applied rather than published. The person who keeps a district model running through monsoon, patches it when data formats change, and explains outputs to a doctor who does not trust machines is doing work as demanding as frontier research for a fraction of the recognition. There is also talent the frontier does not produce at all: the person who holds both technical knowledge and domain judgment to know whether a model is right for the wrong reasons.

India has the depth to build both a champion and a floor, if talent is distributed rather than drained entirely. For countries without India’s depth, the answer is coalition: a regional institute with permanent positions, rotation programs, shared infrastructure. The coalition does not solve the brain drain. It provides a destination closer to home, embedded in a mission the frontier does not offer: building something for the people you grew up with, in the language your mother speaks.