Apply on Employer Site

Voltry · 3 hours ago

Federated Learning & Privacy Research Associate

United States

Contract

Remote

New Grad, Entry Level

Voltry is advancing research in collaboration with the National Renewable Energy Laboratory, focusing on federated learning and differential privacy for power quality forecasting. The role involves implementing and validating a FedProx-based architecture while ensuring privacy protections under various attack scenarios.

Power Grid

Responsibilities

Implement and validate FedProx federation protocol across ESIF laboratory sites

Configure hierarchical topology with 5–50 virtual sites for scalability testing

Implement differential privacy (ε = 4.0) with Gaussian noise on gradients

Conduct gradient inversion attack testing to validate privacy protections

Benchmark federated model accuracy against single-site baseline

Contribute examples to TensorFlow Federated documentation

Co-author methodology sections for IEEE Transactions on Smart Grid submission

Qualification

Federated learning frameworksDifferential privacyPython programmingDeep learning frameworksPublished researchGradient inversion attacksInterest in energy applications