Senior Program Manager, Windows AI
Microsoft · Applied Sciences Group · Belgrade
Working across model, silicon and feature teams to get models from research into the experiences that ship. Currently focused on multi-modal input.
- My day-to-day is making sure the right models reach the features that depend on them — the Settings agent, Video Super Resolution, and the speech-to-text and text-to-speech models among them.
- Those models ship as part of the public Windows AI API surface, so the audience is not only Windows features but any third-party app building on them.
- A lot of the work is matching silicon-optimized models to feature requirements that keep moving, which means staying close to four or five partner teams at once.
- Getting Mu, the small language model behind the agent in Windows Settings, integrated into the experience that ships it was my piece of that work.