About the job
Work Location:
The position requires working from our offices located in Seattle. On-site requirements may differ based on your specific role and team. If you have any inquiries regarding the on-site work arrangements for this position, please reach out to your recruiter.
*Compensation varies based on your current degree enrollment.
Research Intern, Undergrad: $94,320 (Annual)
Research Intern, Masters: $106,110 (Annual)
Research Intern, PhD: $140,000 (Annual)
About You:
The Allen Institute (Ai2) is on the lookout for enthusiastic, early-career researchers to join our team for full-time research internships. Ideal candidates will possess a keen interest in pushing the boundaries of large language model (LLM) based agents and their corresponding environments. We welcome applications from international candidates, and competitive salaries along with visa sponsorship are available.
Internship Details:
- Duration: 12 weeks
- Start Date: Flexible
- Candidate Background: A PhD candidate or a master/undergraduate student with a strong research foundation.
To be considered for our Spring/Summer 2026 internship programs, please submit your application by December 15, 2025, at 11:00 PM Pacific Time.
About Us:
At Ai2, we are passionate about developing large language models that can operate autonomously in real-world settings to meet user objectives. We are pioneering the creation of open agentic LLMs, capable of planning, reasoning, utilizing tools and APIs, coding, and interacting with complex, dynamic environments to safely and reliably solve long-term tasks. This initiative aligns with Ai2’s broader mission to enhance the capability, trustworthiness, and societal benefits of AI systems while also transparently sharing artifacts and effective methodologies for building such systems.
We invite ambitious and driven interns to contribute to our research aimed at advancing the science of training and evaluating general-purpose agentic models. Our projects systematically investigate how data synthesis, training objectives, and learning conditions influence the development of generalizable and reliable agentic behaviors, and how these behaviors can be meaningfully and reproducibly assessed. You will collaborate with a team exploring essential questions in this domain, such as:
- What data, training strategies, and environments yield models that generalize across various agentic tasks?

