Real-Time Decision Systems
Professional work
- Python
- Model serving
- Applied ML
Bring a model into the live workflow
I rebuilt an existing modeling approach as a service that could use fresh inputs when a decision was needed. I owned the work from training and preprocessing through the serving application.
That required attention to both model behavior and the code around it. I worked on input validation, efficient preprocessing, and monitoring so the service could be operated after release.
The result supported automated operational decisions. The project shaped how I think about platform engineering today because it exposed the practical needs of the people who build and run models.