About

Hi! I’m Suze and I’m a junior/coterm at Stanford University, where I study Computer Science (Systems/AI), dance with the Viennese Ball, help lead Stanford ACM and am involved with research. I grew up in the Netherlands and moved to the U.S. after high school for university. I’m a big fan of (Earl Grey) tea, sudokus, consuming baked goods and cheung fun and social dance!

I get excited about natural language processing and machine learning and am currently interested in interactions between model behavior, training data and internals and the science of language models (i.e. scaling laws, emergent capabilities). I lead a research project on how language models (fail to) conceptualize current time under Chris Potts and Diyi Yang and am involved in research on eval scaling laws and emergent capabilities under Tatsunori Hashimoto.

Previously, I worked on a distributed LLM serving system in the MAST group, was a SWE intern at Optiver, competed in ICPC and was an undergraduate teaching assistant for Stanford’s intro CS courses.

Research

I’m still gaining my footing in research –mostly learning, attemps at experimenting and trying not to break too much

	Time in Language Models How do language models reason about time? Do they have an internal "current" time? How is the concept time derived from their training data? Suze van Adrichem, Aditi Bhaskar -- Advised by Jing Huang , Professor Diyi Yang and Christopher Potts Github / Poster
	Predicting Emergent Capabilities [Work in Progress...] Christine Ye, Suze van Adrichem, Neil Band, Luke Bailey -- Advised by Tatsunori Hashimoto

Projects

	Transformer Language Model (CS336) Reimplemented BPE tokenizer, AdamW optimizer, FlashAttention, DDP, ZeRO Stage 1 sharding. Created custom training data curated from Common Crawl. Implemented reinforcement learning method GRPO. 1st in Scaling Laws Competition @ CS336 Suze van Adrichem
	SUZITI Filesystem: FAT32 Filesystem For The Rasberrry Pi Featuring custom I2C, interactive graphical display and 'laptop' enclosure Suze van Adrichem, Aditi Bhaskar Github / Demo
	BAS Buddy: AI-Agent for Climate Initiative Matching Agentic system with custom database to match companies with climate initiatives Best Prototyping, Best Use of GenAI w/ InterSystems Vector Search @ Treehacks Suze van Adrichem, Alice Heiman, Bubble Yu Github / Devpost / Demo

Experience

Optiver, Amsterdam, the Netherlands Software Engineer Intern, Execution Speed and Success, Summer 2025
Stanford NLP Group, Stanford CA, United States Student Researcher, Science of Language Models (Interpretability, Scaling Laws), February 2025-Present
Multi-scale Architectures & Systems Team + CURIS, Stanford CA, United States Student Researcher, Machine Learning Systems, March 2024-June 2025
International Collegiate Programming Contest (ICPC), Stanford CA, United States Team Competitor, September 2024-Present 1st place D2 California regional, 2nd place Northwest Pacific D2 regional (2024)
Stanford ACM, Stanford CA, United States Co-President, May 2025-Present Financial Officer, May 2024-May 2025
CS198 + Stanford Computer Science, Stanford CA, United States Teaching Assistant (Section Leader), January 2024-September 2024

Misc

Last updated: February 16, 2026