Smart Lamp Project & Story

In 2023, I joined a startup team as CTO during the XbotPark winter camp, responsible for end-to-end architecture design of AI products.

Our first attempt was smart sleep earbuds that could automatically pause phone audio/video after the user fell asleep. Due to limited sensor accuracy, the project was put on hold.

We then pivoted to a smart wake-up lamp using mmWave sensing + ML classification to detect sleep stages and wake users in light-sleep / REM. After finishing the core function, I pushed for an MVP-first plan and staged iterations. But with realistic timing & resource risks, the roadmap kept expanding (scope creep).

To secure funding and validate demand, we attended an exhibition in Shenzhen. Worried about scope delays, I temporarily took over as project lead: I completed functional testing, coordinated mechanical & supply chain optimization to accelerate mass production after exposure. The response exceeded expectations at the trade show, and later at a developer conference (GDC), we accumulated many potential users.

However, a request to build a full ecosystem came soon after. Considering the risk of continued scope expansion and the lack of a timely validation & mass-production plan, I decided to exit in July 2024, even though the MVP was ready.

Looking back, the hardest part was balancing exploration vs. delivery. What matters is turning tech demos into shippable products with a clear path to PMF and scale.

Demo Video

Gallery