Mert Cemri

Ph.D. Student, EECS, UC Berkeley

• Email
• Google Scholar
• LinkedIn
• X (Twitter)
• CV

About

I am a Ph.D. student in the Berkeley AI Research (BAIR) and Sky Computing Labs at UC Berkeley, advised by Ion Stoica, Kannan Ramchandran, and Alex Dimakis.

My research focuses on efficient machine learning and agentic systems, including multi-agent large language model (LLM) systems, LLM-driven discovery and optimization, test-time scaling, and discrete generative models.

I am also a part of the Berkeley Laboratory for Information and System Sciences (BLISS) and an organizer of the BLISS Seminar. Feel free to reach out if you would like to give a talk.

News

• Sep 2026 Four papers accepted to NeurIPS 2026, including an oral for Inductive Deductive Synthesis.

Selected publications

1. Fantastic Adaptive Taxonomies and How to Use Them arXiv M. Cemri*, A. Cojocaru*, M. Pan, S. Liu, S. Agarwal, A. Krentsel, J. Tang, K. Ramchandran, J. E. Gonzalez, M. Zaharia, A. Dimakis, I. Stoica ICML Workshop on Failure Modes in Agentic AI (FAGEN), 2026 · Best Paper Award
2. SPECS: Faster Test-Time Scaling through Speculative Drafts and Dynamic Switching arXiv M. Cemri*, N. Rajaraman*, R. Tiwari*, X. Liu, K. Keutzer, I. Stoica, K. Ramchandran, A. Beirami, Z. Sun Advances in Neural Information Processing Systems (NeurIPS), 2026
3. Inductive Deductive Synthesis: Enabling AI to Generate Formally Verified Systems arXiv S. Agarwal, A. Krentsel, S. Liu, M. Cemri, A. Cheng, R. Meng, T. Pfister, C.-L. Li, S. Ratnasamy, A. Parameswaran, M. Zaharia, I. Stoica, M. Lesani Advances in Neural Information Processing Systems (NeurIPS), 2026 · Oral
4. Spectral Feedback for Test-Time Alignment of Protein Diffusion Models Paper S. Dickman, M. Cemri, L. Butler, K. Ramchandran Advances in Neural Information Processing Systems (NeurIPS), 2026
5. AdaEvolve: Adaptive LLM Driven Zeroth-Order Optimization arXiv M. Cemri*, S. Agrawal*, A. Gupta, S. Liu, A. Cheng, Q. Mang, A. Naren, L. E. Erdogan, K. Sen, M. Zaharia, A. Dimakis, I. Stoica arXiv preprint, 2026
6. EvoX: Meta-Evolution for Automated Discovery arXivCode S. Liu*, S. Agarwal*, M. Maheswaran, M. Cemri, Z. Li, Q. Mang, A. Naren, E. Boneh, A. Cheng, M. Z. Pan, A. Du, K. Keutzer, A. Cheung, A. G. Dimakis, K. Sen, M. Zaharia, I. Stoica Conference on Language Modeling (COLM), 2026
7. SkyDiscover: A Flexible, Adaptive Framework for AI-Driven Scientific and Algorithmic Discovery Code S. Liu, M. Cemri, S. Agarwal, A. Krentsel, A. Naren, Q. Mang, Z. Li, A. Gupta, M. Maheswaran, A. Cheng, M. Pan, E. Boneh, K. Ramchandran, K. Sen, M. Zaharia, A. G. Dimakis, I. Stoica ACM Conference on AI and Agentic Systems (CAIS), 2026 · Industry Spotlight
8. Why Do Multi-Agent LLM Systems Fail? arXiv M. Cemri*, M. Z. Pan*, S. Yang*, L. A. Agrawal, B. Chopra, R. Tiwari, K. Keutzer, A. Parameswaran, D. Klein, K. Ramchandran, M. Zaharia, J. E. Gonzalez, I. Stoica Advances in Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track, 2025 · Spotlight
9. DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning arXiv H. Bai*, Y. Zhou*, M. Cemri, J. Pan, A. Suhr, S. Levine, A. Kumar Advances in Neural Information Processing Systems (NeurIPS), 2024
10. Asynchronous Social Learning M. Cemri, V. Bordignon, M. Kayaalp, V. Shumovskaia, A. H. Sayed IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2023
11. Discovering Influencers in Opinion Formation over Social Graphs arXiv V. Shumovskaia, M. Kayaalp, M. Cemri, A. H. Sayed IEEE Open Journal of Signal Processing, 2023

* denotes equal contribution. For the full list, see Google Scholar or my CV.

Experience

• Google, Systems Research Group · Student Researcher Sunnyvale, CA · April 2026 – present · AI for systems: LLM-driven agents for systems performance research and synthesis of formally verified systems components
• Apple, AI/ML · Machine Learning Intern Cupertino, CA · Summer 2025 · Agentic post-training for foundation models
• EPFL, Adaptive Systems Lab · Research Intern with Ali H. Sayed Lausanne, Switzerland · Distributed optimization and social learning
• Bilkent University · Research with Tolga Çukur and Aykut Koç Ankara, Turkey · Neural methods for natural language and graph-structured data

Awards

• Oral, NeurIPS 2026 (Inductive Deductive Synthesis)2026
• Best Paper Award, ICML Workshop on Failure Modes in Agentic AI (FAGEN)2026
• Spotlight, NeurIPS Datasets and Benchmarks Track2025
• EECS Departmental Fellowship, UC Berkeley2023–2024
• Research Excellence Award and Academic Excellence Award, Bilkent University2023

Contact

The best way to reach me is by email at cemri@berkeley.edu.