Production AI feature: smart search or ticket deflection
Ship a real AI feature for a software product, with monitoring, a security review and a demo.
The brief
A software product's users cannot find answers, so they open support tickets. Add an AI feature that fixes this, either smart search over the help content or a ticket-deflection assistant that answers before a ticket is filed, and ship it the way a product team would. Good looks like: a deployed feature, measured quality and cost, a prompt-injection review, monitoring dashboards and a recorded demo for a panel.
What you will build
- Choose the feature and define success metrics
- Build it with retrieval over the help content
- Evaluate quality on a test set before release
- Review security against the OWASP Top 10 for LLM applications
- Deploy with monitoring of quality, latency and cost
- Present the result, trade-offs and next steps to a panel
Free, open data to build it with
- Shopping Queries Dataset (ESCI) ↗ Amazon Science (GitHub)
- Customer support tickets ↗ Hugging Face
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.

