Research Intern

Compresr•San Francisco, CA

About The Position

We are building state-of-the-art context compression, aiming to become the 'Cloudflare for LLMs'. Our mission is to embed a compression layer into most LLM pipelines by default. We are a team of ex-EPFL MSc/PhDs who started by publishing papers, then joined YC and began generating revenue by assisting companies in reducing their LLM expenses. We operate our business like a research lab, forming hypotheses, discarding ineffective ones, and focusing on successful strategies. This internship offers competitive compensation, all necessary resources (GPUs, subscriptions, OpenAI/Anthropic credits), significant responsibility, and a fast-paced learning environment with technically proficient colleagues. There is a possibility of a full-time offer based on performance. However, we do not offer hands-on supervision; guidance is high-level, and interns are expected to own their work. Projects are not pre-defined due to our early-stage, customer-and-market-driven approach, requiring interns to navigate multiple directions. After a brief onboarding, interns will tackle challenging, customer-facing, and time-sensitive problems alongside the team. This is not a typical internship; it's a demanding environment designed for rapid growth and skill development.

Requirements

Love research, read papers and hack on new repos for fun
Comfortable training ML models/transformers
Comfortable doing independent applied research
Excellent Claude Code (or similar) user
Highly ambitious, ready for high-intensity YC startup culture, self-motivated
Strong communicator
Fast response time
Team player

Nice To Haves

LLM research experience, shown through publications, open-source contributions, or personal projects
BSc or MSc in CS/DS, math, or physics
Startup or research internship experience (industry or academic)