I Now

I am a medical student at Washington University in Saint Louis School of Medicine. My research applies statistical and computational approaches to surgery, and to medicine at large.

II Road So Far

I graduated from Johns Hopkins University with a BS in Computer Science in May 2019 and an MSE in Applied Mathematics & Statistics in December 2019. While there I was fortunate to be advised by Suchi Saria and Avanti Athreya. During my gap year I worked as an Assistant Research Engineer at my alma mater. I was a part of the NeuroData lab where I worked with Carey Priebe on statistical graph inference and with Joshua Vogelstein on causal inference from observational health data.

I started medical school at WashU in 2021. In 2024, I took a research year during which I was the CNS AI fellow at NYU's OLAB under Dr. Eric Oermann, co-advised by Dr. Eric Leuthardt at WashU. During that time I trained the first neurosurgical chatbot that can view imaging — CNS-Obsidian — which was evaluated in a real-world randomized trial.

See CNS-Obsidian read imaging in three real-world neurosurgical scenarios:
Subarachnoid hemorrhage Spinal epidural hematoma Acute ischemic stroke
III Research

My work spans clinical artificial intelligence, computational safety, and connectome statistics. Selected papers below.

Clinical Artificial Intelligence Models

2026 CNS-Obsidian: A Neurosurgical Vision-Language Model Built From Scientific PublicationsA. Alyakin, J. Stryker, D. A. Alber, et al.Neurosurgery 2025 CNS-CLIP: Transforming a Neurosurgical Journal into a Multimodal Medical ModelA. Alyakin, D. Kurland, D. Alber, et al.Neurosurgery 2025 MedMobile: A Mobile-sized Language Model with Clinical CapabilitiesK. Vishwanath, J. Stryker, A. Alyakin, et al.BMJ Digital Health & AI 2025 Lang1: Generalist Foundation Models Are Not Clinical Enough for Hospital OperationsL. Y. Jiang, A. Chen, X. Han, …, A. Alyakin, et al.preprint

Computational Safety

2026 Generalist Large Language Models Outperform Clinical Tools on Medical BenchmarksK. Vishwanath, A. Alyakin, M. Ghosh, et al.Nature Medicine (accepted) 2026 LLM-assisted Systematic Review of Large Language Models in Clinical MedicineS. F. Chen, A. Alyakin, A. Seas, et al.Nature Medicine 2025 Medical Large Language Models are Vulnerable to Data-poisoning AttacksD. A. Alber, Z. Yang, A. Alyakin, et al.Nature Medicine 2025 Pitfalls of Multiple-Choice Questions in Generative AI and Medical EducationS. Singh, A. Alyakin, D. A. Alber, et al.Scientific Reports 2025 Evaluating the Performance and Fragility of Large Language Models on Neurosurgical Board-like QuestionsK. Vishwanath, A. Alyakin, M. Ghosh, et al.Neurosurgery 2025 Medical Large Language Models are Easily DistractedK. Vishwanath, A. Alyakin, D. A. Alber, et al.preprint

Connectome & Graph-Based Statistics

2024 Correcting a Nonparametric Two-sample Graph Hypothesis Test for Graphs with Different Numbers of Vertices, with Applications to ConnectomicsA. A. Alyakin, J. Agterberg, H. S. Helm, C. E. PriebeApplied Network Science 2021 Valid Two-Sample Graph Testing via Optimal Transport Procrustes and Multiscale Graph Correlation, with Applications in ConnectomicsJ. Chung, B. Varjavand, J. Arroyo, A. Alyakin, et al.Stat 2021 An Analysis of Euclidean vs. Graph-Based Framing for Bilingual Lexicon Induction from Word Embedding SpacesK. Marchisio, Y. Park, A. Saad-Eldin, A. Alyakin, et al.Findings of the ACL: EMNLP

Commentary

2025 "All You Need" is Not All You Need for a Paper Title: On the Origins of a Scientific MemeA. Alyakinpreprint