Vladislav Akavets

Trained tree-based models for fraud prediction: Achieved a 92% approval rate and <0.45% fraud rate initially; improved performance to 93% approval rate and <0.3% fraud rate.
Performed EDA and feature engineering using blockchain data to enhance model performance for crypto merchants.

Trained tree-based ML models (XGBoost, Random Forest) for churn prediction and upsell. Developed custom scoring metrics for those.
Built a pipeline to move paper document information into digital form, using Tesseract text detection on top of picture quality improvement layers.
Optimized SQL queries in PL/SQL and built visualization in Tableau.

Grade: 4.12

Activities and societies: Chess Club

Graduated top of the class, coursework focused on ML and Math. Led two projects in Computer Vision Object Detection. Authored a paper on Reinforcement Learning/Optimal Control.

Grade: 3.96

Activities and societies: Indoor Soccer Club, Chess Club

Graduated from the Honors program. Double degree in Computer Science and Mathematics. Won Rasor-Bareis-Gordon Math competition twice and ACM programming contest with a team of 3 once. Coursework focused in ML and Computer Vision.