Role Overview This internship is focused on ML research mixed with applied ML and software engineering. Topic of the internship is around trust and safety in LLMs such as: - prompt injections - safety alignment and guardrails - privacy and confidentiality - interpretability Other topics closer aligned with the student's research may be considered. Expected Background: ML - Research experience - Ability to pose relevant questions, find answers in existing literature or plan and execute experiments. - Ability to read and write research papers and technical articles. - PhD student or student at any level with research experience and peer-reviewed published work. - Machine learning tools: pytorch, huggingface, transformers, datasets. - Applied deep learning and LLM experience. - Training and evaluating deep models. - (nice to have) finetuning LLMs, multi-modal LLMs. - (nice to have) Familiarity with ML[NLP,LLM,Vision] interpretability methods, sparse autoencoders, linear probes. Expected Background: Software Engineering - Development environments and tools: - unix, git, basic clouds usage on AWS and/or GCP - jupyter - Programming: - python - (nice to have) “programming languages well-roundedness” - experience in statically-typed and functional languages Compensation & Benefits - Market aligned compensation for interns in the bay area. Requirements - Must be authorized to work in the USA or must be able to obtain CPT (Curricular Practical Training) approval from host university. - The internship is expected to last at least 3 months and intersect most of the period starting from January, 2027 through end of March 2027.
Summary
This internship focuses on ML research in trust and safety for LLMs, including prompt injections, safety alignment, and interpretability. Candidates need research experience with peer-reviewed publications, proficiency in PyTorch, Hugging Face, and Python, and authorization to work in the US or CPT approval.
Responsibilities
ML research mixed with applied ML and software engineering; topics include prompt injections, safety alignment and guardrails, privacy and confidentiality, interpretability.
Qualifications
PhD student or student at any level with research experience and peer-reviewed published work; ability to pose relevant questions, find answers in existing literature or plan and execute experiments; ability to read and write research papers and technical articles.
Education requirements
PhD student or student at any level with research experience and peer-reviewed published work.
Experience requirements
Research experience; ability to pose relevant questions, find answers in existing literature or plan and execute experiments; ability to read and write research papers and technical articles.