Fellow in AI​/ML Scientific , Foundation Models, and AI Systems

School: Faculty of Arts and Sciences

The Kempner Institute at Harvard University seeks early‑career researchers to help shape the future of AI as Kempner AI Fellows. We are looking for candidates with strong foundations in modern machine learning and the ambition to advance foundation models, agentic systems, and new AI approaches for high‑impact scientific applications.

We seek candidates with strong technical preparation in modern AI/ML, a demonstrated record of research accomplishment, and an interest in contributing to ambitious research at the frontier of machine learning and science. Areas of particular interest include:

• foundation model training, evaluation, and adaptation
• agentic workflows, tool‑using models, and AI systems
• bespoke scientific applications of AI/ML, including in the life sciences
• alternative architectures and systems‑level approaches to modern AI

AI Fellows will work closely with one another and with Kempner faculty, researchers, and students on foundational machine learning and domain‑informed scientific applications. The position is particularly well‑suited to candidates eager to deepen their technical skills in foundation models and agentic systems while applying those methods to important scientific and technical problems.

• Fellows will conduct research under the direction of a Kempner Institute investigator.
• Fellows are appointed for a one‑year term; reappointment may be possible for up to three consecutive years.
• Due to the importance of in‑person mentoring, this position is based on campus, full‑time, at Harvard University. Remote work for this position is not possible.

• Bachelor's or master's degree in computer science, statistics, electrical engineering, applied mathematics, computational biology/neurobiology, or a related quantitative field required by the expected start date
• Strong technical background in modern AI/ML, including deep learning and hands‑on experience with frameworks such as PyTorch or JAX
• Demonstrated research productivity, including publications in venues such as ICML, ICLR, NeurIPS, or similar, and/or substantial open‑source research contributions
• Demonstrated experience implementing, training, evaluating, or fine‑tuning modern machine learning models
• Strong programming skills in Python and experience building and maintaining research code
• Demonstrated ability to use modern AI‑assisted and agentic coding tools effectively, such as Claude Code, Codex, or similar systems, in research and development workflows
• Ability to work effectively in a collaborative research environment and communicate technical work clearly

• Experience with foundation model training, post‑training, adaptation, or evaluation
• Experience with agentic workflows, tool‑using models, retrieval systems, or related AI systems
• Experience with large‑scale datasets, distributed training, or high‑performance computing environments
• Interest in scientific applications of AI/ML, including the life sciences
• Interest in alternative architectures and systems‑level approaches to AI

Harvard University is committed equal opportunity and non‑discrimination. We seek talent from all parts of society and the world, and we strive to ensure everyone at Harvard thrives. Our differences help our community advance Harvard's academic purposes. Harvard has an equal employment opportunity policy that outlines our commitment to prohibiting discrimination on the basis of race, sex, ethnicity, color, national origin, religion, disability, or any other characteristic protected by law or identified in the university's non‑discrimination policy.

Harvard's equal employment opportunity policy and non‑discrimination policy help all community members participate fully in work and campus life free from harassment and discrimination.

Expected annual salary is $54,600 for candidates holding a bachelor's degree and $60,060 for candidates holding a master's degree. Salary will be commensurate with qualifications and experience.

Molly Marshall

Contact Email:
Kempner Institute