Spring Intern — Agentic AI Discovery & Prototyping

Genentech; Daly City (CA); On-site; USD 59,000 - 84,000; Full time

Benefits offered by this job

Paid internship

Stipend provided

On-site in South San Francisco

Job summary

Genentech is seeking a Spring Intern to advance Agentic AI Discovery & Prototyping within the Data Science & Machine Learning team. You will prototype agent workflows, benchmark frameworks, and ensure safety and regulatory alignment while integrating LLMs with internal APIs.

The internship runs for 12 weeks, full-time, with a possible extension. A location-based stipend supports costs; work is on-site in South San Francisco, CA.

Qualifications

• Currently enrolled in a graduate degree program (e.g., CS, DS, AI/ML, or related field).
• Foundational proficiency in Python and version control.
• Understanding of LLMs and Generative AI patterns; prompt engineering.

Responsibilities

• Build, test, and refine AI agent workflows for interactive QA, search, and automation.
• Design a framework to benchmark LLM outputs, agent reliability, tool calling, and reasoning accuracy.
• Implement guardrails for safe, compliant agent outputs in healthcare contexts.
• Develop clean Python code to integrate LLMs with internal APIs and data sources.
• Collaborate with engineers, data scientists, and stakeholders to gather requirements and iterate on feedback.
• Present project findings, architectures, and metrics to the team at internship end.

Skills

Python

Git

LLMs

AI safety

Prompt engineering

Agentic workflows

Education

Master's Degree (enrolled student)

PhD (enrolled student)

Tools

LangChain

LlamaIndex

APIs

Job description

Genentech is seeking a Spring Intern to advance Agentic AI Discovery & Prototyping within the Data Science & Machine Learning team. You will prototype agent workflows, benchmark frameworks, and ensure safety and regulatory alignment while integrating LLMs with internal APIs.

The internship runs for 12 weeks, full-time, with a possible extension. A location-based stipend supports costs; work is on-site in South San Francisco, CA.