ML Researcher — Interpretability & Next-Gen Architectures

Tilde Research; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job description

Tilde Research is a moonshot AI lab advancing mechanistic interpretability, new architectures, and pretraining science. We build foundational understanding of models to advance the frontier of intelligence.

About the role:

As a ML Researcher, you will develop innovative techniques to deeply understand how large AI models work—and use those insights to make them better. You'll work on training, analyzing, and evaluating cutting-edge models, collaborating closely with a team of researchers and engineers to advance interpretability as a tool for improving performance and control.

What you might work on:

• Designing, prototyping, and optimizing novel model architectures

• Investigating why models or specific components behave the way they do—and how to improve them

• Curating targeted datasets to elicit and instill specific behaviors or capabilities in models

• Collaborating on papers, blog posts, and open-source tools to share insights with the broader community

You're a good fit if you:

• Have experience in deep learning or related research areas

• Have deep technical expertise in some subfield(s) of modern ML, e.g. architecture, optimizers, RL, learning dynamics, etc.. This can include through:

  • Strong publication record

  • Thoughtful technical blog posts

• Have experience working with large-scale pre/post-training infrastructure for experimentation

• Communicate clearly and effectively, both verbally and in writing

• Can come up with and evaluate research ideas

• Are able to learn quickly