My Time at Freddie Mac
A reflection on my experience working at Freddie Mac, where I contributed to innovative tech solutions in the finance industry.
Deployed and shipped TensorBoard feature from non-prod to production environment, followed by creation of tests and demos of feature on self-service platform. Tested custom-developed and pre-existing machine learning models using PyTorch, to validate and demonstrate capabilities of TensorBoard feature for internal clients. Used linear regression, neural networks, CNNs, RNNs, clustering, and dimensionality reduction. These techniques were used to configure data visualization across TensorBoard dashboards including: Scalars, Graphs, Distributions, and Histograms primarily. Tests for Audio, text, Image and Projector dashboards were developed as well. Created permanent documentation for Data Scientists and ML engineers regarding: Tensorboard, configuring it's usage on our self-service DS platform, PyTorch utilities specific for tensorboard, and more. Demoed findings and results for data department.
Wrote Data Preprocessing, Model Training, Model Deployment, and End-to-End ML Workflow Data Science Pipelines on KubeFlow pipelines tool to demonstrate updated features in a video for platform users.
Worked with a team of interns to outline implementation of a future technology for Freddie Mac’s Technology intern summer challenge. Developed initial idea and outlined chosen concept, continuing by defining plans for development, implementation and adoption with a team of 10+ talented interns throughout the summer. Presented concept to executives at Freddie Mac.
Delievered tensorboard feature for dozens of active users within Freddie Mac's FDP (freddie data platform)