# Bounty. AI

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Kavodel Akinola](https://cerebralvalley.ai/u/kavodel)
- **GitHub:** https://github.com/Kavii101/bouncy-ai-tts/tree/feat/interview-copilot
- **Demo video:** https://www.youtube.com/watch?v=X4gbouH9SE8
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4588

Bounty Online Co-pilot is an AI interview and presentation copilot that turns a job description and resume into a complete, evidence-based performance, and then helps you deliver it live.

    The problem: talented candidates and analysts lose opportunities not on knowledge, but on preparation and delivery. They struggle to connect their experience to what a role actually requires, they practise without honest feedback, and
 they freeze when presenting live.

    How it works: Setup - paste a job description and resume. Brief - the AI extracts an evidence-linked brief showing requirements, proof from your experience, gaps, and likely questions; nothing proceeds until you approve it. Practice -
 answer questions and receive rubric feedback that flags unsupported claims. SQL Lab - generate queries that pass through a read-only safety gate that rejects destructive SQL. Live - with consent, the copilot listens, transcribes in real
time, and streams concise suggested answers to a separate Answers teleprompter window. Present and Subtitles - a cue-card prompter and live captions for screen-sharing. A Listen panel also converts reports and text into natural-voice audi
o for accessibility.

    Built as a full-stack TypeScript app (React, Express, tRPC, Drizzle) behind feature flags, with consent-gated capture, hashed practice answers, redacted logging, 139 passing automated tests, and a Docker and GHCR deployment pipeline.
Every claim in the product is grounded in the user's own materials: the copilot helps you say what you can prove.

## More from Meta Global AI Developer Hackathon

- [Ferrum](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4585)
- [IncidentLens](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4586)
- [Witness Warehouse Solution](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4587)
- [Assemvio](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4589)
- [Lianne Baron](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4590)
- [walmartpotato](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4591)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4588. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt