About the role#
This internship focuses on a mix of machine learning research and applied software engineering. The work centers on trust and safety in large language models, covering areas such as prompt injections, safety alignment, guardrails, privacy, and interpretability. Candidates may also propose topics aligned with their existing research interests.
What you'll do#
- Conduct research on LLM trust and safety.
- Pose relevant questions, survey existing literature, and execute experiments.
- Read and write research papers and technical articles.
- Train and evaluate deep learning models.
- Utilize tools including PyTorch, HuggingFace, Transformers, and datasets.
- Apply software engineering best practices within Unix, Git, and cloud environments like AWS or GCP.
What you'll need#
- Currently pursuing a degree in Computer Science, Data Science, Mathematics, Computer Engineering, or Electrical Engineering.
- Demonstrated research experience.
- Proficiency in Python and Jupyter notebooks.
- Experience with applied deep learning and LLMs.
- Ability to work on-site in Sunnyvale, California.
- Authorization to work in the USA or ability to obtain CPT approval.
- Familiarity with fine-tuning LLMs, multi-modal models, interpretability methods, sparse autoencoders, or linear probes is a plus.
- Experience with statically-typed or functional programming languages is a plus.
Location & details#
- Location: Sunnyvale, CA
- Term: Fall 2026
- Modality: On-site
- Employment: Full-time, paid internship
How to get in at Realm Labs
Intern Insider's per-company playbook.
Securing an internship at Realm Labs requires timing because the team is small and roles fill quickly. Intern Insider sends an instant alert the moment a role matching your target is published anywhere, so you apply among the first. Early applicants get seen before the pile grows, which is critical for a lean organization. You should also focus on finding the right point of contact to demonstrate genuine interest. Intern Insider surfaces the recruiters behind the company's roles so you can reach out directly to ask about the role or a referral. This approach materially improves response rates compared to standard submissions.
