AI Differential Privacy Engineer

Telvion Labs

Remote · $160k – $200k / year · Full-time · Posted 5 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

Synthetic data is only as useful as the privacy guarantees it provides. If a synthetic dataset leaks information about the individuals it was generated from, it is not private — it is just noisy. Differential privacy gives you formal, provable guarantees about what a synthetic dataset can and cannot reveal, but implementing DP correctly is subtle and implementing it efficiently enough for real-world use is genuinely hard. We've built the infrastructure for generating synthetic tabular, time series, and text data at scale. The privacy math works. What we need now is someone who can push the utility-privacy frontier: better noise calibration that sacrifices less utility per unit of privacy budget, membership inference testing frameworks that give clients evidence of privacy protection, and tooling that helps data scientists work with DP concepts without needing a PhD in statistics. This is a research-engineering role. We publish our methods and engage with the academic DP community. You'll need to read and implement recent papers, and you may write some.

Fully remote.

Responsibilities

Improve noise calibration mechanisms to reduce utility loss per privacy budget unit
Build membership inference attack frameworks for privacy guarantee testing
Develop tooling for data scientists to work with DP concepts intuitively
Implement DP-SGD and related techniques for differentially private ML training
Benchmark privacy-utility trade-offs against academic baselines
Engage with the DP research community through publications and open source

Requirements

5+ years in ML engineering with privacy, security, or statistics specialization
Deep understanding of differential privacy theory and implementation
Experience with DP-SGD, the moments accountant, or related techniques
Strong Python and PyTorch skills
Publications or significant contributions to privacy-preserving ML research preferred