Privacy Safe Person Re-ID System

I am assembling an end-to-end solution that can reliably follow the same individual across multiple cameras while keeping their facial identity hidden and their personal data footprints minimal. The goal is to balance privacy and utility, not to sacrifice one for the other.

What the build must cover
• Facial obfuscation that can be reversed only by authorized parties
• Robust multi-camera tracking and re-identification pipelines
• Quantitative privacy-assurance metrics that prove the strength of the protection

I will look to you to design the core AI/CV architecture, code and train the models, wire them into a working prototype, and document measurable results on standard datasets. If you are comfortable blending deep-learning vision models, LLM-based policy logic, and solid programming practices, you will fit right in.

To be considered, simply outline the experience you have that is directly relevant to privacy-preserving computer vision or large-scale multi-camera tracking. Links to demos or publications are welcome.

For background on the wider initiative, feel free to browse the current projects here: https://sites.google.com/view/kaai-vision917/home.

If you’d like updates on future consultancy openings, or part-time roles, subscribe to our channel: https://youtube.com/@kaai-visions?si=xMNj_MugNtN-Su1a.

I’m keen to collaborate with experts who can turn this concept into a demonstrable, privacy-respecting system.