Posey placed 1st in the final round of WELTARE Spring Sprint 1.1.4: Startup track, hosted by AI Tinkerers — Paris. The same round also recognized Radar (Abdulmalik) as 1st place for an AI-powered job-search platform—congratulations to that team as well. Full listings and copy are on the official results page: https://paris.aitinkerers.org/hackathons/h_kiyKmulxWFk/results.
What judges said about Posey
Posey perfectly bridges experimental pose-estimation with a scalable, venture-ready architecture.
Its explainable AI core, mobile-first technical strategy, and clear market positioning make it a standout professional-grade startup contender.
That feedback lines up with what we set out to prove during the sprint: real-time pose estimation is not a demo trick—it can sit inside a product story that is credible for users, mentors, and anyone asking “why this, why now?”
Posey in one paragraph
Posey is a judgment-free AI fitness coach: real-time pose detection, voice coaching, and guidance aimed especially at people who find traditional gyms or trainers hard to access. If you want the full product breakdown, see the Posey project page →.
What we leaned into during the sprint
Pose estimation that ships
From experiments to something you can run, measure, and show—latency, stability, and real sessions.
Explainability where it matters
Users should understand why the coach said what it said—not a black box scolding their form.
Mobile-first execution
A native, on-device feel for coaching: the interface is the product, not a wrapper around a slide deck.
The event page also summarized the team behind Posey as combining strengths across LLMs, 5G security research, and full-stack Go / TypeScript (including experience as an NLP engineer, a PhD security researcher, and a GE HealthCare software engineer)—the kind of cross-disciplinary mix you need when “AI coach” touches models, client performance, and production architecture.
Finalists we shared the stage with
The preliminary round advanced strong builders alongside Posey, including teams such as Inteply, Ship mode, Radar, and Omar & Alex (names as shown on the results page). That density of execution is exactly why we keep showing up to AI Tinkerers Paris.
Acknowledgements
Thank you to AI Tinkerers Paris for running the program, to WELTARE for backing this edition, and to the judges and peers who pushed every demo to be clearer than the last.
The Posey team
Phuong Nhi Chau
Hanh Thi-Hong Tran
Tien Ngoc Viet
Tan Nhat Linh LEThank you for supporting Tihado and Posey.