ML / AI Research Intern at Oleria in Bellevue, WA

Summary: Oleria Security is focused on providing adaptive and autonomous identity security solutions. As an ML / AI Research Intern, you will lead a 12-week project to research, prototype, and validate a machine learning-driven access intelligence engine, aimed at improving access provisioning for organizations. Responsibilities: - You'll own this project end-to-end -- from research and data pipeline through working prototype and final presentation. Core work includes: - ML research and prototyping - Research, implement, and compare unsupervised learning approaches (e.g., k-means, hierarchical clustering, graph-based methods) to identify peer-group cohorts from employee attributes and entitlement data - Generative AI integration - Connect an LLM layer that translates ML outputs into human-readable cluster names and recommendation rationales that non-technical administrators can act on - Synthetic data pipeline - Design and generate realistic datasets with employee attributes, entitlement histories, and usage signals to support training, evaluation, and offline testing - Extensible framework - Build a modular, well-documented codebase designed to serve as the foundation for future access intelligence work beyond this internship - Technical report - Document your methodology, experiments, results, and recommendations in a written report that informs the production roadmap - Final presentation - Demo the working prototype and present key learnings to the engineering team and leadership Required Qualifications: - Strong undergraduate or recent graduate in Computer Science with foundational knowledge of AI/ML and generative AI concepts - You've used LLMs to build something -- in a class, a hackathon, a side project. You understand where they work well and where they don't, and you can talk through the tradeoffs - Solid understanding of supervised and unsupervised learning concepts, feature engineering, and model evaluation. You know how to design a test set and report results honestly - Some exposure to experiment tracking, reproducibility, or structuring ML code for reuse -- whether from coursework, a project, or self-study - You can explain your work clearly to both technical and non-technical audiences, in writing and in conversation - You hold yourself to a high bar, push through ambiguity, and don't consider something done until it's actually good - You can take a scoped project, manage your own progress with periodic check-ins, and see it through to completion Preferred Qualifications: - Some exposure to identity, access management, or security concepts -- through coursework, reading, or personal interest - Familiarity with graph-based or network-based ML concepts - Coursework or projects touching NLP or large language models - Experience writing technical documentation or research reports Required Skills: Machine learning, Generative AI, Unsupervised learning, Supervised learning, Feature engineering, Model evaluation, Experiment tracking, ML code structuring, Identity, access management, Graph-based machine learning, Natural language processing, Drive, Self-directed Benefits: Mentorship from builders, Real ownership, Production-grade experience, Durable impact, A culture that brings out your best