Staff Research Scientist, Intern - PhD Agentic AI

About the role

Visa’s Internship Program offers a 12‑week immersive experience where you’ll work on impactful projects that advance Visa’s mission. As a Visa intern, you’ll build meaningful connections, strengthen your communication and business skills, and gain hands‑on experience in a dynamic, global environment. You’ll have access to interactive workshops and learning sessions designed to build expertise, expand industry knowledge, and elevate your professional skills. You’ll not only learn, you’ll contribute, collaborate, and innovate. Beyond professional development, enjoy social events that foster community and connection. The program culminates in a final presentation, where you’ll share your project achievements and recommendations with Visa leaders—a chance to demonstrate impact and growth.

Responsibilities

• Conduct research to improve AI agent reliability, reduce hallucinations, and advance safety, security, and ethical alignment.
• Design techniques to enhance AI agent efficiency, such as model compression, quantization, fine‑tuning strategies, and inference optimization.
• Develop and evaluate secure access control and guardrail systems that promote responsible AI interactions.
• Investigate multi‑agent architectures for secure, coordinated, and efficient collaboration.
• Build tools and frameworks to benchmark agent safety, robustness, and performance.
• Prototype and experiment to translate research ideas into practical AI capabilities.
• Contribute to research publications, patents, and technical reports where appropriate.

Requirements

• Students pursuing a Ph.D. with a graduation date in December 2027 or later.
• Digital fluency, including the ability to work with emerging technologies such as Generative AI tools (e.g., ChatGPT, Microsoft Copilot) to support everyday work.
• Position performed in an office setting; requires sitting and standing at a desk, communicating in person and by telephone, and frequently operating standard office equipment such as telephones and computers.

Qualifications

• Strong research record demonstrated by publications, patents, conference presentations, or technical contributions.
• Research experience in AI agents, AI safety, security, alignment, trustworthy AI, or reliable LLM systems.
• Experience with multi‑agent systems, agent frameworks, planning, or evaluation techniques.
• Knowledge of techniques for reducing hallucinations, enhancing factuality, enabling guardrails, enforcing policies, or building secure AI systems.
• Experience applying efficiency methods such as compression, quantization, fine‑tuning, or inference optimization.
• Proficiency in programming languages such as Python, C++, Java, or similar.
• Strong communication, collaboration, and analytical problem‑solving skills.
• Ability to innovate, adapt to ambiguity, and tackle complex research challenges.
• Curiosity and continuous learning.

Skills

Programming proficiency (Python, C++, Java, etc.), research methodology, AI agent design, safety and alignment techniques, efficiency