
Rajat Monga, CVP of AI Frameworks at Microsoft, details his career trajectory across the frontier of AI infrastructure, from co-founding Google Brain and TensorFlow to navigating startup pivots and current edge AI development. Building core infrastructure requires a "refounding" mindset, where teams must periodically reassess whether existing systems can support future goals or if a clean-slate approach is necessary. Successful adoption of new technology hinges on aligning tools with human habits and organizational incentives, as even technically superior solutions often fail if they disrupt existing power dynamics or workflows. Monga emphasizes that while large-scale cloud models drive innovation, the future of AI lies in optimizing for latency, privacy, and cost on local edge devices, enabling practical applications ranging from agricultural automation to advanced robotics.
Sign in to continue reading, translating and more.
Continue