Machine learningJun — Aug 2024PM Intern, Team Lead
Wine Quality ML
Leading a team of 3 to ship a predictive model people could use.
+15%
accuracy
−20%
compute time
+30%
engagement
01
The problem
Existing wine-quality prediction was slow and inaccurate, and hard for non-experts to use.
02
How I approached it
- 1.Researched users and compared supervised learning algorithms.
- 2.Led a team of 3 through model development.
- 3.Oversaw a friendly interface for real-world use.
03
What shipped
Predictive quality modelUser-friendly interface
Next project
Komu →