Working Backwards: Defining the Core Flow
Instead of jumping straight into UI, I focused on two key questions:
1️⃣ What does it look like to train a dataset right away?
2️⃣ How can we surface the right insights at the right time?
2️⃣ How can we surface the right insights at the right time?
To answer this, I worked with the team to:
Outline requirements and validate assumptions.
Sketch potential friction points from the user’s perspective.
Test messaging through multiple homepage variations to simplify the value proposition.
Aligning Epics to Design Priorities
Once we had a clearer vision, we mapped Epics that aligned both technical and design priorities:
🔹 Fast-Track Model Training – Users needed a streamlined way to upload, clean, and train data without unnecessary steps.
🔹 Data Visibility for Decision-Makers – GMs and coaches needed insights, not just raw numbers.
🔹 Smart Defaults for Non-Experts – Removing friction by suggesting configurations based on best practices.
🔹 Data Visibility for Decision-Makers – GMs and coaches needed insights, not just raw numbers.
🔹 Smart Defaults for Non-Experts – Removing friction by suggesting configurations based on best practices.
Execution: Designing an ML Project for Simplicity
I focused on turning a highly technical workflow into an intuitive experience. This included:
✅ Clear onboarding that guided users through dataset preparation.
✅ A simplified training dashboard that removed unnecessary toggles and jargon.
✅ Real-time insights tailored for sports professionals, not just data scientists.
✅ A simplified training dashboard that removed unnecessary toggles and jargon.
✅ Real-time insights tailored for sports professionals, not just data scientists.
Results & Impact
Lowered the barrier to entry for non-technical users.
Increased usability of ML tools for sports teams, making trained data more actionable.
Validated key UX assumptions through iterative testing, ensuring alignment with user needs.
By working backwards from the ideal experience, we turned machine learning into a tool, not a hurdle—giving teams more power over their data without needing a PhD in AI. 🚀
Final Takeaway
By working backwards from the ideal user experience, we turned machine learning from a complex tool into a strategic advantage. The final product empowered sports teams to harness the power of AI without needing deep technical expertise—making data-driven decision-making faster, easier, and more effective. 🚀
