Performance & Experience Team

Led the design and development of novel privacy-preserving machine learning solutions in collaboration with a top financial institution as part of an undergraduate thesis.

Outperformed tranditionally trained model by 6.5% using primary performance metrics in a public competition.

Publishing our unique solution, leveraging the open-source platform Flower and PyTorch to implement a Split NeuralNetwork for classification.

Conducted data analysis on public large-scale customer transaction data using Pandas, while transforming and storing datain a PostgreSQL database.

Designed and implemented two deep neural network architectures, DnCNN and UNet, for removing noise frommedical imaging scans using PyTorch.

Analyzed the characteristics of the noise distribution in the internal dataset and used this analysis to augment thedataset, resulting in a 35-fold increase in size.

Trained the models and achieved a 9.7% and 11.8% increase in mean signal-to-noise ratio for the DnCNN and UNetmodels, respectively, without loss of tissue structure.

Optimized deployment of applications to 1000+ stores by 30% by developing a web app to simply confi gure deployments,using Python/Flask, JavaScript, HTML, and PostgreSQL.

Developed a Flask API that effi ciently manages client requests, enabling the creation and modification of deployment information stored in a PostgreSQL database.

Integrated and managed user authentication to ensure secure access to sensitive deployment information.

Reduced time spent coding by 30% by completely automatating Terraform configuration code for provisioning cloud resources.

Designed and developed GitLab CI/CD pipelines, leveraging Docker and Kubernetes, to deploy containerize microserviceson a GKE cluster.

Led bi-weekly Scrum meetings for a team of 15 people, using Jira for ticket management.