Skip to content

Capstone: a multimodal assistant with a latency budget

Build a small assistant that takes a picture or a voice note and answers, measure where the time and tokens go, and write a cost and latency report. Deliverables, rubric and where to go next.

150 min·Verified 2026-11Hands-on labPro+100 XP
Share

You can now turn images and audio into numbers, estimate their cost, and serve the models that read and write them. This capstone builds one small assistant and measures it honestly.

Pro content

Keep reading with Inference Academy Pro

Unlock this full lab, plus every advanced lesson, hands-on lab, the larger playground models and completion certificates.