Dr Amirmohammad Farzaneh
I am a Career Development Lecturer in Computer Science at Pembroke College and a Postdoctoral Research Associate at Northeastern University London. My research asks how we can make AI systems not only accurate, but reliable enough to act autonomously in uncertain and changing environments. I work at the intersection of machine learning, statistical inference, information theory and networked systems, developing methods that provide meaningful performance and risk guarantees for agentic AI.
My route to Pembroke began with a BSc in Electrical Engineering at Sharif University of Technology, followed by a DPhil in Engineering Science at Oxford. During my DPhil, I studied the complexity and compression of network topologies through an information-theoretic lens. My subsequent research at King’s College London and Northeastern University London has focused on reliable machine learning, including conformal prediction, post-selection evaluation, counterfactual inference and uncertainty-aware autonomous systems.
I joined Pembroke in 2026 after several years of teaching engineering, mathematics and computer science across Oxford. At Pembroke, I teach Computer Science and support students through Oxford’s tutorial system. I particularly enjoy helping students connect abstract ideas to the behaviour of real computational systems, and encouraging them to ask not only whether an algorithm works, but when and why it can be trusted.
What I love most about my subject is its combination of elegant theory and practical consequence. Questions about uncertainty, information and decision-making can lead to beautiful mathematics while also shaping how AI systems operate in communications, networks and other settings where dependable decisions matter.
Detailed biography
Dr Amirmohammad Farzaneh is a Career Development Lecturer in Computer Science at Pembroke College, a Postdoctoral Research Associate at Northeastern University London and a Non-Stipendiary Lecturer in Engineering Science at Oriel College. He completed his DPhil in Engineering Science at the University of Oxford in 2024. His thesis developed an information- theoretic approach to understanding the complexity and compression of network topologies. Before coming to Oxford, he completed a BSc in Electrical Engineering at Sharif University of Technology, graduating top of his programme.
His research spans reliable machine learning, agentic AI, conformal prediction, post-selection evaluation, counterfactual inference, information theory and networked systems. His work has appeared in venues including NeurIPS, ICML, IEEE Signal Processing Letters, IEEE ISIT, EUSIPCO and the Journal of Complex Networks.
Dr Farzaneh is an Associate Fellow of the Higher Education Academy. He has taught engineering mathematics, control theory, computer engineering, information engineering, engineering ethics, mathematics and computer science. He has also supervised and mentored doctoral and master’s students.
His academic service includes organising and presenting a tutorial on statistically valid hyperpa- rameter selection at NeurIPS 2025, serving on the AAAI programme committee and reviewing for leading machine-learning and information-theory conferences. He received an ICML 2025 Reviewer Gold Medal. While at Mansfield College, he served for two years as President of the Middle Common Room, having previously served as Vice-President.
His distinctions include runner-up in the University of Oxford Three Minute Thesis competition, Mansfield College Student of the Year and a Moogsoft research award supporting his doctoral work.
Selected recent publications
A. Farzaneh and O. Simeone, “Statistically Valid Post-Training Hyperparameter Selection: From Tuning to Guarantees,” monograph, 2026.
A. Farzaneh and O. Simeone, “Post-Selection Distributional Model Evaluation,” NeurIPS, 2026.
A. Farzaneh, S. D’Oro and O. Simeone, “Should I Have Expressed a Different Intent? Counterfactual Generation for LLM-Based Autonomous Control,” ICML, 2026.
A. Farzaneh and O. Simeone, “Multi-Objective Hyperparameter Selection via Hypothesis Testing on Reliability Graphs,” NeurIPS, 2025.
A. Farzaneh, S. Park and O. Simeone, “Quantile Learn-Then-Test: Quantile-Based Risk Control for Hyperparameter Optimization,” IEEE Signal Processing Letters, vol. 31, pp. 3044– 3048, 2024.
A fuller publication list is available on Google Scholar
Dr Amirmohammad Farzaneh
I am a Career Development Lecturer in Computer Science at Pembroke College and a Postdoctoral Research Associate at Northeastern University London. My research asks how we can make AI systems not only accurate, but reliable enough to act autonomously in uncertain and changing environments. I work at the intersection of machine learning, statistical inference, information theory and networked systems, developing methods that provide meaningful performance and risk guarantees for agentic AI.
My route to Pembroke began with a BSc in Electrical Engineering at Sharif University of Technology, followed by a DPhil in Engineering Science at Oxford. During my DPhil, I studied the complexity and compression of network topologies through an information-theoretic lens. My subsequent research at King’s College London and Northeastern University London has focused on reliable machine learning, including conformal prediction, post-selection evaluation, counterfactual inference and uncertainty-aware autonomous systems.
I joined Pembroke in 2026 after several years of teaching engineering, mathematics and computer science across Oxford. At Pembroke, I teach Computer Science and support students through Oxford’s tutorial system. I particularly enjoy helping students connect abstract ideas to the behaviour of real computational systems, and encouraging them to ask not only whether an algorithm works, but when and why it can be trusted.
What I love most about my subject is its combination of elegant theory and practical consequence. Questions about uncertainty, information and decision-making can lead to beautiful mathematics while also shaping how AI systems operate in communications, networks and other settings where dependable decisions matter.
Detailed biography
Dr Amirmohammad Farzaneh is a Career Development Lecturer in Computer Science at Pembroke College, a Postdoctoral Research Associate at Northeastern University London and a Non-Stipendiary Lecturer in Engineering Science at Oriel College. He completed his DPhil in Engineering Science at the University of Oxford in 2024. His thesis developed an information- theoretic approach to understanding the complexity and compression of network topologies. Before coming to Oxford, he completed a BSc in Electrical Engineering at Sharif University of Technology, graduating top of his programme.
His research spans reliable machine learning, agentic AI, conformal prediction, post-selection evaluation, counterfactual inference, information theory and networked systems. His work has appeared in venues including NeurIPS, ICML, IEEE Signal Processing Letters, IEEE ISIT, EUSIPCO and the Journal of Complex Networks.
Dr Farzaneh is an Associate Fellow of the Higher Education Academy. He has taught engineering mathematics, control theory, computer engineering, information engineering, engineering ethics, mathematics and computer science. He has also supervised and mentored doctoral and master’s students.
His academic service includes organising and presenting a tutorial on statistically valid hyperpa- rameter selection at NeurIPS 2025, serving on the AAAI programme committee and reviewing for leading machine-learning and information-theory conferences. He received an ICML 2025 Reviewer Gold Medal. While at Mansfield College, he served for two years as President of the Middle Common Room, having previously served as Vice-President.
His distinctions include runner-up in the University of Oxford Three Minute Thesis competition, Mansfield College Student of the Year and a Moogsoft research award supporting his doctoral work.
Selected recent publications
A. Farzaneh and O. Simeone, “Statistically Valid Post-Training Hyperparameter Selection: From Tuning to Guarantees,” monograph, 2026.
A. Farzaneh and O. Simeone, “Post-Selection Distributional Model Evaluation,” NeurIPS, 2026.
A. Farzaneh, S. D’Oro and O. Simeone, “Should I Have Expressed a Different Intent? Counterfactual Generation for LLM-Based Autonomous Control,” ICML, 2026.
A. Farzaneh and O. Simeone, “Multi-Objective Hyperparameter Selection via Hypothesis Testing on Reliability Graphs,” NeurIPS, 2025.
A. Farzaneh, S. Park and O. Simeone, “Quantile Learn-Then-Test: Quantile-Based Risk Control for Hyperparameter Optimization,” IEEE Signal Processing Letters, vol. 31, pp. 3044– 3048, 2024.
A fuller publication list is available on Google Scholar