Sr. Researcher, Hybrid AI Security Team Lead

Candidate must be willing to work onsite three days a week in our Tel-Aviv, Israel office. 

Lenovo Digital Trust Lab is seeking a hands-on AI Evaluation and Hardening Team Lead to join our AI Security team. This role leads the team responsible for evaluating and improving the security, privacy, safety, and reliability of AI models and applications. 

You will own the technical roadmap and delivery of Lenovo’s AI Evaluation Framework, combining applied research, platform engineering, and adversarial testing. Working closely with research, product, and governance teams, you will drive policy-aligned assessments and turn evaluation findings into better training data, stronger system prompts, and safer models. 

Job Responsibilities: 

• Lead and develop a multidisciplinary team of engineers and researchers, balancing research, platform delivery, and operational needs. 

• Own the Evaluation Framework roadmap, architecture, delivery, and production reliability. 

• Advance evaluation methodologies, adversarial testing, and automated red teaming across models and agentic applications. 

• Translate governance and product requirements into measurable assessments, scoring, and actionable reports. 

• Drive evaluation-to-hardening feedback loops, including training data generation, model fine-tuning, and system prompt improvements. 

• Partner with product and model teams to integrate evaluation and hardening into development and release workflows. 

Minimum Requirements: 

• 7+ years of experience in AI/ML, security research, or software engineering, including 2+ years leading technical teams. 

• BSc or MSc in computer science or a related field, or equivalent practical experience. 

• Hands-on experience with LLMs, model evaluation, and model training or fine-tuning. 

• Strong Python skills and experience delivering production software or research platforms. 

• Experience designing experiments, datasets, and metrics to assess model behavior and validate improvements. 

• Demonstrated ability to own technical roadmaps, mentor team members, and deliver across research and engineering. 

• Strong communication skills and the ability to translate complex findings into clear priorities. 

Preferred Requirements 

Experience in one or more of the following: 

• AI Security: Adversarial ML, AI red teaming, prompt injection, jailbreaks, or agentic security. 

• Evaluation Methods: LLM-as-judge, benchmark design, automated adversarial testing, or multimodal evaluation. 

• Model Hardening: Adversarial training, safety fine-tuning, preference optimization, or instruction hardening. 

• AI Governance: Translating security, privacy, and safety requirements into technical tests and evidence. 

• Evaluation Platforms: Scalable evaluation services, model pipeline integrations, or testing of agentic workflows.