Recently at Meta Reality Labs Research working on future experiences for AI wearables with advanced sensors beyond microphones and cameras. Previously led the design of visual search on the Amazon Shopping app (now called Lens) where I helped launch multiple generations of the product.
Prototyped an AI agent using the Ray-Ban Meta glasses to enable parents of newborns to easily log essential baby activities like diaper changes & sleep schedules, without pulling out the phone from the pocket.
This experiment shows a path towards a world where job-oriented agents could replace apps on the next computing platform.
A short article summarizing my thoughts about a new relationship model between users, OS and agents on wearables - one that is different from what we have with apps on the current computing platform.
I hope to spark more conversations as a new generation of wearables comes up in the next few years.
For the first time we have a promising solution for personalizing the UI for a long tail of scenarios thanks to AI's ability to generate UI on-the-go.
I chose to focus on a smaller screen to strike the balance between personalization and predictability.
Created proactive agent experiences on a two-part AI wearable constellation.
Established new shell models for future AR devices that handle nondeterministic input and output during human x AI interactions.
My attempt to communicate more system and model states subtly and intuitively to users at the time when nondeterministic outputs are still novel.
A brief walk through of multiple generations of the product experience during my time there, and the reason behind the key pivot in each generation.
