ML Research Scientist - Differential Privacy

DP-DFL

Our client, a software company specializing in deploying large language models (LLMs) for enterprises and refining these models using sensitive data, recently disclosed raising $15.1 million in a Series A funding round. This funding was co‑led by investment firms Canapi Ventures and Nexus Venture Partners. Additionally, Formus Capital and Soma Capital participated in this funding round, contributing to a total raised amount of $19.3 million for our client.

• Lead and manage an ML privacy vertical focused on specific domains and optimization strategies (e.g., differential privacy, privacy‑enhancing algorithms, anonymization).
• Collaborate closely with the engineering team to implement practical applications of your developed algorithms, catering to our clients' needs.
• Contribute significantly to the creation of research papers, patents, and presentations by integrating your vertical's work with other team members' research contributions.

• While our primary focus involves federated, distributed, and privacy‑centric learning, extensive experience in federated learning (FL) is not mandatory. However, we require:
  • Profound expertise in privacy‑preserving ML methodologies.
  • Hands‑on experience in training Differentially Private models utilizing frameworks such as Opacus, OpenDP, etc.
  • Extensive practical knowledge in deploying various Large Language Model (LLM) models and architectures in real‑world scenarios. Proficiency in leading end‑to‑end projects is expected.
  • Desirable:
    Domain expertise in understanding attacks against ML models, especially in Computer Vision (CV) or Large Language Models (LLMs).
• Adaptability and flexibility are crucial. In the academic and startup realms, staying updated with new findings in the field may prompt sudden shifts in focus. Therefore, the ability to swiftly learn, implement, and extend state‑of‑the‑art research is imperative.