Playment's Jarvis & Pixel Design System

Playment is a fully managed data labeling platform that generates training data for computer vision models at scale using crowdsourcing. The motto is to empower companies in the Autonomous Vehicle, Drones, Mapping, and similar spaces with high-precision annotation services. We are a young company backed by Y-Combinator and SAIF Partners; we have helped the likes of Nio, Didi Chuxing, the University of Washington, Nuro, Drive.ai, and many more to fuel their vision of Autonomous Vehicles.

At the core of its platform is a trained workforce of 300,000+ users (Players/Annotators) managed by their human intelligence experts. This workforce builds annotation tasks on the training data and delivers quality results. As part of this, I was tasked to design Jarvis, our product for orchestrating crowdsourcing of data, as well as our design system and annotation tool - semantic segmentation.

Designing enterprise software was an entirely new experience for me. There were so many constraints, complexities, stakeholders, rules of governance, features, functionalities, technicalities, personas, and hundreds of use cases to consider. It was a challenge to come up with an ideal experience that was able to support many of the existing features while providing routes to establish new user behaviors. This tension between abstraction and customization is not unique to this industry, but it’s a problem that is faced in many different industries.

Finally, Challenges and Problem-Solving are essential parts of the day-to-day operations of a hyper-growth startup. The efficiency of project management and the product directly impact the potential revenues. To address this, we are aiming to fill the void with our product, which is designed to streamline the process and make it easier to manage, while also ensuring the highest quality results.