FlowSense | Devpost

Inspiration

We were inspired by a simple observation: the same 20 minutes can feel completely different depending on what you are doing. Time seems to fly when you are deeply engaged, but drags when you are distracted or stressed. We wanted to build a system that measures not only objective time, but also subjective time—how time actually feels to the user—and use that to better understand focus, flow, and performance.

What it does

FlowSense is a wearable cognitive-sensing system that uses smart glasses to measure focus, flow state, and subjective time perception during tasks like studying, working, or gaming. By combining eye tracking, EEG, and heart-rate sensing, it identifies when users are deeply engaged, when attention drops, and how their perception of time changes throughout a session. After the task is finished, users can review a clear report showing STR (Subjective Time Rate), flow intervals, and session-level insights.

How we built it

We designed FlowSense as a multimodal system with both hardware and software components. On the hardware side, we conceptualized smart glasses integrating inward-facing eye tracking, EEG contact points, and heart-rate sensing. On the software side, we designed a data pipeline that collects sensor streams, preprocesses them into synchronized time windows, and feeds them into a transformer-based model for multimodal fusion. We also designed a companion web app that displays post-session reports, including flow timelines, STR curves, and long-term analytics.

Challenges we ran into

One major challenge was defining subjective time in a measurable way. Unlike heart rate or gaze, subjective time is not directly observable, so we had to think carefully about how to represent it as a model output. Another challenge was integrating multiple sensing modalities into a product form factor that still feels wearable and realistic. We also had to balance scientific ambition with product usability, making sure the system could produce insights that feel understandable and useful to everyday users.

Accomplishments that we're proud of

We are proud that FlowSense goes beyond traditional productivity tracking by introducing subjective time as a new dimension of personal analytics. We built a strong end-to-end concept: a wearable sensing product, a multimodal AI pipeline, and a user-facing analytics platform. We are also proud of shaping a technically ambitious idea into something that feels like a plausible real product rather than just a research concept.

What we learned

We learned that focus is not just about productivity—it is also about perception. Measuring cognitive state requires combining signals that each tell only part of the story. We also learned how important system design is when building AI products: sensors, model architecture, UX, and interpretation all need to work together. Most importantly, we learned that invisible mental states can become much more actionable when translated into simple, intuitive feedback.

What's next for FlowSense

Next, we want to move from concept to prototype by building a working sensing stack, testing with real user sessions, and validating whether our model can reliably estimate flow and subjective time. We also want to improve personalization, since cognitive signals vary greatly from person to person. On the product side, we plan to expand the companion app with richer comparisons across tasks, environments, and times of day, helping users discover when and how they do their best work.