Bio
I am a Marie Skłodowska-Curie Doctoral Fellow at Imperial College London, advised by Prof. Alessandra Russo. My research interests lie in generalist agents, world models, and AI safety. Previously, I worked at Microsoft AI on post-training Microsoft 365 Copilot for tool use and reasoning.
Before pursuing a PhD in AI, I co-founded Athina AI (YC W23), an LLM observability startup, and worked in machine learning roles at Turo, BMW, and Amazon. Please feel free to reach out if you share similar research interests or would like to discuss transitioning from the start-up world to a PhD.
Publications & Preprints
VITA: Zero-Shot Value Functions via Test-Time Adaptation of Vision-Language Models.
C. Ziakas*, A. Russo.
arXiv, 2025 (under review). [Paper]
Red-Bandit: Test-Time Adaptation for LLM Red-Teaming via Bandit-Guided LoRA Experts.
C. Ziakas*, N. Loo, N. Jain, A. Russo.
arXiv, 2025 (under review). [Paper]
Towards Shutdownable Agents via Stochastic Choice.
E. Thornley*, A. Roman*, C. Ziakas*, L. Ho, L. Thomson.
NeurIPS 2025 Aligning RL Experimentalists and Theorists Workshop, 2025. [Paper]
Test-Time Adaptation for Generalizable Task Progress Estimation.
C. Ziakas*, A. Russo.
ICML 2025 Test-Time Adaptation Workshop, 2025. [Paper]
BackboneLearn: A Library for Scaling Mixed-Integer Optimization-Based Machine Learning.
V. Digalakis*, C. Ziakas*.
arXiv, 2023. [Paper]
Scalable Econometrics on Big Data — The Logistic Regression on Spark.
A. Ouattara*, M. Bulté, W. J. Lin, P. Scholl, B. Veit, C. Ziakas, et al.
arXiv, 2021. [Paper]
Neural Temporal Point Processes for Capacity Forecasting.
C. Ziakas.
MSc Thesis, Technical University of Munich (advisor: Prof. Stephan Günnemann), 2021.
Implementation of Decision Trees for Data Streams on Spark.
C. Ziakas.
Diploma Thesis, Technical University of Crete (advisor: Prof. Minos Garofalakis), 2018. [Paper]
* denotes first authorship.