Chapter

Agentic & multimodal systems

Tool use and function calling, action grounding, streaming voice and vision, agent memory, and multi-agent orchestration.

Tech lead

Real-time AI manager and coach

Real-time voice agent that manages and coaches the user through their work, pulls up resources when needed, and checks the deliverable as it forms.

Major contributor

Tool-using agents in simulated apps

Simulation runtime that can stand up any software tool and run an agent or a user through it, with specific understanding checkpoints baked into every task.

Sole author

Multi-persona AI manager and team (Do Mode)

Multi-persona conversation engine that puts a fictional company's team around the user, with the manager and the right teammate stepping in for the moment.

Co-contributor

Learn Mode, sibling flow to Do Mode

Guided understanding flow where the manager walks the user through what they need to grasp for the deliverable, sequenced with pre-generated videos and an optional cowork session at the end.

Sole author

Voice-first first-time onboarding

Voice agent that builds a user's profile (goals, prior experience, how they frame problems) from a short first conversation, with no forms.

Sole author

Presentation defense with live Q&A

Real-time voice flow where the user delivers their work to an AI audience, fields live follow-up questions tied to what they actually said, and receives a structured verdict at the end.