Electrification took decades to show up in measured productivity. Now, intelligence is evolving faster than the world can put it to work.

When intelligence becomes abundant, what becomes the bottleneck to productivity? Permissioned access and execution. Optimizing across both delivers real-world outcomes with automatically improving economics. We go as deep into the stack and as far into the last mile as outcomes demand: for inference execution, continual learning informs model orchestration and custom kernels on heterogeneous hardware; where a licence or access is the last-mile constraint, we own the outcome end to end, starting with investment management. One feedback loop runs through every layer.

This is a call for India's best talent. We are hiring the strongest people in AI systems infrastructure and post-training. Our team already includes people who chose Mumbai after working in California, New York, London, Dubai and Singapore. We would rather leave a seat empty than fill it below our bar.

We hire for evidence of exceptional ability and agency. You arrive with a defining strength and build range around it. Roles are starting positions, and accountability travels with you. We choose consequential, under-solved problems, hold ourselves to the world's best, and count progress only when something real has changed. If that sounds like you, write to talent@optima.ai and send the work you're proudest of.