Internship - Private and Secure Agentic AI

Mitsubishi Electric Research Laboratories (MERL) is seeking a highly motivated and qualified PhD student for an internship focused on privacy-preserving and secure agentic AI systems. The role involves collaborating with researchers to develop technologies that enhance privacy and security, with the goal of publishing results at leading AI research conferences.

Responsibilities

• Conduct research on privacy-preserving and secure agentic AI systems
• Collaborate with MERL researchers on developing novel fundamental technologies that enhance the privacy and security of agentic systems that employ foundation models, such as LLMs/VLMs
• Aim to publish new scientific results at top-tier AI research conferences

Skills

• Currently pursuing a PhD in Computer Science, Electrical Engineering, or a related field
• Strong background in machine learning, LLMs/VLMs, foundation models, and agentic AI systems
• Research experience with private and/or secure foundation models (e.g., differential privacy, adversarial inputs, jailbreaking attacks/defenses, prompt injection)
• Proficiency in Python and deep learning frameworks, such as PyTorch and Hugging Face tools
• Proven publication record in top-tier ML/AI research conferences

Benefits

• Relocation stipend
• Covered travel to and from MERL
• Monthly Charlie Card for local commuting
• Weekly social gatherings
• Professional development opportunities, including research talks by both internal and external speakers
• Health insurance coverage

Company Overview

Company H1B Sponsorship