I'm a founder working on on-device multimodal AI.
Before that I spent five years at Google Brain and Google DeepMind, where I led research on multimodal language models. I was the tech lead for multimodal LaMDA, the precursor to Gemini, worked on Gemini's multimodal capabilities, and launched compact language models that run on-device on Pixel phones. Earlier I built real-time machine learning systems for self-driving cars at Cruise.
I have a PhD from Vanderbilt University, where I worked on creating AI models for cyber-physical systems, and a bachelor's in electrical engineering from the University of Tehran.
Selected publications
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
G. Comanici, E. Bieber, M. Schaekermann, I. Pasupat, N. Sachdeva, I. Dhillon, and others. arXiv, 2025. - Gemini: A Family of Highly Capable Multimodal Models
Gemini Team, Google. arXiv, 2023. - LaMDA: Language Models for Dialog Applications
R. Thoppilan, D. De Freitas, J. Hall, N. Shazeer, and others. arXiv, 2022. - Adversarial Regression for Detecting Attacks in Cyber-Physical Systems
A. Ghafouri, Y. Vorobeychik, X. Koutsoukos. IJCAI, 2018. - A Game-Theoretic Approach for Selecting Optimal Time-Dependent Thresholds for Anomaly Detection
A. Ghafouri, A. Laszka, W. Abbas, Y. Vorobeychik, X. Koutsoukos. Autonomous Agents and Multi-Agent Systems, 2019. - Optimal Thresholds for Anomaly-Based Intrusion Detection in Dynamical Environments
A. Ghafouri, W. Abbas, A. Laszka, Y. Vorobeychik, X. Koutsoukos. GameSec, 2016.
The full list is on Google Scholar.