Group Product Manager, Frontier Model Defense

DeepMind Technologies Limited; Mountain View (CA); Sur place; USD 300 000 - 333 000; Plein temps

Avantages offerts par ce poste

Equity

Bonus target 25%

Benefits

Résumé du poste

Google DeepMind seeks a Product Manager to own the frontier model protection strategy, guiding end-to-end development from conception to launch while balancing defense with model utility and developer experience.

You will lead cross‑functional teams, set governance, and shape safety benchmarks with researchers, engineers, and external partners in a fast-paced, highly innovative environment.

Qualifications

• Bachelor's degree or equivalent practical experience.
• 10 years of experience in product management or a related technical role.
• 5 years of experience taking technical products from conception to launch.

Responsabilités

• Define the multi-year product goal and strategic roadmap for frontier model protection and anti-distillation defenses across Google DeepMind.
• Establish policies and runtime guardrails to balance rigorous IP defense against developer friction, model utility, and commercial growth objectives.
• Develop quantitative evaluation frameworks and telemetry to measure defense efficacy and continuously validate complex trade‑off decisions.
• Partner with researchers and engineers to translate breakthroughs in watermarking and query analysis into globally scalable, production-grade defenses.
• Secure executive alignment and lead technical engagement with peer labs and AI Safety Institutes to establish shared evaluation benchmarks.

Connaissances

Product management

Technical product development

Roadmap planning

Cross-functional collaboration

Formation

Bachelor's degree or equivalent

Description du poste

Google DeepMind seeks a Product Manager to own the frontier model protection strategy, guiding end-to-end development from conception to launch while balancing defense with model utility and developer experience.

You will lead cross‑functional teams, set governance, and shape safety benchmarks with researchers, engineers, and external partners in a fast-paced, highly innovative environment.