Data

About the work

Smashspeed uses computer-vision models — built on TrackNet v3 — to track the shuttlecock frame by frame and turn it into a smash speed. Those models are only as good as the data behind them, and that's where you come in.

What you'll do

• Annotate shuttlecock positions across video frames to build our TrackNet v3 training datasets.
• Follow — and help refine — our labeling guidelines so the data stays consistent and high-quality.
• Flag tricky edge cases like motion blur, occlusion, and unusual angles.
• Help with dataset QA and, if you're interested, basic model evaluation.
• Pitch in on other random tasks around the startup — we're small and everyone wears a few hats.

You might be a good fit if you

• Are precise, patient, and consistent — you notice the details others miss.
• Are curious about machine learning and computer vision.
• Are reliable and can hit agreed turnaround times.
• Bonus: some Python, or experience with annotation tools or CV datasets.

Details

You'll be working closely with a tight-knit group of ambitious builders. It's a supportive, high-energy environment where everyone is genuinely productive, ideas turn into shipped work fast, and you get the real, unfiltered experience of building a startup from the ground up.

This role is focused on hands-on ML experience. You'll watch your work train real models and learn how a computer-vision data pipeline actually runs. Strong contributors can take on more responsibility over time.

We're bringing on several annotators, so apply even if you can only help part-time.