Research Fellowship - Mechanistic Interpretability
- Location
- San Francisco, CA · On-site
- Job type
- Temporary
- Level
- Senior
- Education
- Doctorate
- Listed
- Today
Skills & highlights for this role
- mechanistic interpretability
- reinforcement learning
- PyTorch
- Python
Also useful
- JAX
- sparse autoencoders
- LLM post-training
How well does your resume fit this job?
Upload your resume to see your score for this Vmax role, what matches, what's missing, and your best other matches.
Scored on our server and not saved there. The file stays in this browser so you don't have to upload it again after you sign up.
About the role
LLMs are fantastically powerful and there is a rapidly growing corpus of work devoted to understanding their internal representations and computations. We use the tools of mechanistic interpretability to enhance reinforcement learning by generating intrinsic rewards as a supplement or alternative to downstream human-generated verifiers.
This 3 to 6 month fellowship is for PhD students or equivalent early-career researchers who want to work at the intersection of mechanistic interpretability and reinforcement learning. You will own a focused research project, work closely with Vmax technical staff, and contribute to research publications.
Responsibilities
- Develop mechanistic interpretability methods for understanding internal representations, features, circuits, and computations in language models and agents.
- Investigate how model internals can be used to generate intrinsic rewards, auxiliary objectives, diagnostics, or training signals for reinforcement learning.
- Design and run experiments that test whether interpretability-derived signals improve learning, exploration, generalization, robustness, or sample efficiency.
- Compare internally derived rewards against baselines such as human-generated verifiers, reward models, task-level outcome rewards, and standard intrinsic motivation methods.
- Use techniques such as probing, activation analysis, sparse autoencoders, causal interventions, feature attribution, or representation analysis to study model behavior.
- Analyze failure modes, including reward hacking, spurious features, non-causal correlations, objective misspecification, and overfitting to narrow evaluation distributions.
- Build research code, evaluation harnesses, and experimental infrastructure that make results reproducible and useful to the broader team.
- Communicate research progress clearly through written updates, internal presentations, and final project outputs.
Role Requirements
- Currently enrolled in a PhD program in machine learning, computer science, artificial intelligence, computational neuroscience, mathematics, or a related technical field. Exceptional candidates with equivalent research experience may also be considered.
- Track record of research excellence or strong research promise, demonstrated through publications, preprints, open-source work, technical projects, competitions, or publicly available artifacts.
- Working understanding of reinforcement learning.
- Familiarity with mechanistic interpretability, representation analysis, or empirical methods for understanding neural networks.
- Strong programming ability in Python and experience with at least one major ML framework such as PyTorch or JAX.
- Clear written and verbal communication of technical ideas.
Nice to have
- Experience with LLM post-training methods
- Familiarity with intrinsic motivation, unsupervised RL, auxiliary objectives, representation learning for RL, or curiosity-driven learning.
- Experience with scalable ML experimentation, distributed training, experiment tracking, or reproducible research infrastructure.
- Interest in turning mechanistic understanding into practical training methods, rather than only analyzing models after training.
Role specific location policy
- This role is based in our San Francisco office; for exceptional candidates we are willing to consider a hybrid arrangement
About V max
V max is an applied research lab developing AI capable of open-ended learning. We are building systems to exceed humans in all capacities by optimizing beyond the local maxima of learning from human expertise.
From Vmax's careers page, first listed on Rolev.ai on November 19, 2026. Rolev.ai is not the employer.
How Rolev.ai helps with this application
Rolev.ai is a job search and application tool for US job seekers. For this Research Fellowship - Mechanistic Interpretability role at Vmax, it shows how your resume fits, tailors your resume to the posting, and prepares the application for you to review.
See how you fit
Openings from employer career sites are ranked against your whole resume—and searchable inside Claude, ChatGPT, and Cursor via Rolev.ai MCP.
Tailor without inventing
Every change to your resume is checked against your original. New experience goes in only with your yes—zero fabricated claims.
Draft the full kit
Tailored and Full Match resumes, role-specific cover letters, and answers to employer screening questions are prepared in seconds.
1-Click Extension Autofill
The Chrome extension fills the employer's Greenhouse, Workday, Lever, or Ashby form in 1 click. You review it and press Submit.
Similar roles
All Research Scientist jobs- Research InternMakerMaker · San Francisco, CA
- Research ScientistOpenmind Technologies Inc. · San Francisco, CA
- Senior/Lead Research Scientist, Model EvaluationsThomson Reuters · United States
- Research ScientistParallel · San Francisco or Palo Alto
- ResearcherAka Us · San Francisco, CA
- Staff Research Engineer, Multi-Agent ScalingAnthropic · San Francisco, CA
More at Vmax
Common questions
Where is the Research Fellowship - Mechanistic Interpretability job at Vmax located?
This Research Fellowship - Mechanistic Interpretability position at Vmax is located in San Francisco, CA (On-site). Always check Vmax's application page for specific state eligibility or relocation policies.
What qualifications are required for Research Fellowship - Mechanistic Interpretability at Vmax?
For the Research Fellowship - Mechanistic Interpretability role, Vmax highlights a doctorate; experience with mechanistic interpretability, reinforcement learning, PyTorch, Python. Rolev.ai can score your existing resume against these requirements and show any skill gaps before you apply.
How do I apply for the Research Fellowship - Mechanistic Interpretability position at Vmax?
This listing links directly to Vmax's Greenhouse application page. You can apply directly on Vmax's site, or use Rolev.ai to check your resume fit, generate a zero-hallucination tailored resume and cover letter for this posting, and autofill the Greenhouse application form in 1 click with the Rolev.ai Chrome extension. You always review the form and press Submit yourself.