Two University of Michigan CS PhDs with extensive machine learning background and work experience at Amazon. Two additional UMich undergrad CS Majors contributing.
- Scaled up Machine Learning pipelines: 4600 processors, 35000 GB memory achieving 5-minute execution
- Designed a new Machine Learning pipeline to replace existing prod: AUC perf. increase from 83% to 90%
- Handled 2+ TB data with graphs up to 130 GB (50M nodes, 100M edges) using single-node in-disk scaling