Google Assistant
+ Gemini
Roles: Conversational AI design · Content design · Synthetic data creation
I worked within the Google Assistant team in the areas of content and conversation design for multimodality on wearables.
Some of my responsibilities included assisting with fine tuning Gemini responses, creating synthetic data for training, and designing content to deliver global customer experiences that were helpful, timely, and relevant.
Due to the confidential nature of projects that are bound by an NDA, my work can only be discussed at a higher level.
The team
Conversation/Content Designer · Product Designers · UX Researchers · Product Managers · Engineering
Projects
— PROJECT
ANDROID AUTO PROJECTED
Gemini response framework
Announced Google I/O 2025
Gemini in-dash: for those times when you need an assistant that you can chat with to get information while on the move.
My role involved working a Product Manager to first understand user needs for conversational AI interactions, and then create sample conversations for Gemini that would provide value to the user in every turn.
More will be shared once this gets fully released.
Note: These images are screen grabs from the I/O presentation. Better quality visuals will be added after release.
— PROJECT
WEARABLES
Gemini response framework
Gemini, while being useful, was extremely verbose, and tended to over-explain when asked even simple questions. In addition to this, it also added information that wasn’t relevant to the information being asked.
Sometimes, users just need short, simple answers to their queries due to device and/or environmental considerations, and at such times Gemini’s response can’t be the primary focus of their attention. Additonally, overly verbose responses in audio-only scenarios could become annoying, leading to customer dissatisfaction.
Presented below is evidence of Gemini’s verbosity. Note the overabundance of information.
Efficiently concise –
Only the facts, please
The challenge here was to create a framework that would guide Gemini to tailor its responses to be helpful, concise, and efficient while being informative across different use cases. It would need to be able to respond to the user’s prompt to the best of its ability without going off the rails with its response length.
I collaborated with a product design colleague to create a framework based on actual Gemini responses, along with tailored synthetic data. The ideal outputs lengths and the information they would contain covered the most common use cases, as well as important edge cases.
Note: The actual framework and associated deliverables are covered by an NDA, but this visual demonstrates some of the kind of thinking that informed the creation of the framework.
