Carlo Ciliberto

Carlo Ciliberto

Associate Professor in Machine Learning
Acting Director - AI Centre
Chief AI Officer - Century

Openings

I am always looking for motivated students and researchers to join my group. If you are interested in working with me, please see the information below.

Quick Navigation:

Research Environment

All positions are based at the UCL AI Centre, a world-leading hub for AI research. A vibrant, interdisciplinary community with access to state-of-the-art computational resources and opportunities for high-impact research.

Current PhD Fellowships (Starting 2026)

đź”” PhD Fellowship: Trustworthy and Reliable Agentic AI through Formal Verification

Location: AI Centre, University College London (UCL) Start Date: September 2026

We are seeking an exceptional PhD candidate to join the UCL AI Centre and pioneer research into the formal verification of agentic AI systems. Agentic AI, which combines large language models with planning and tool use, holds immense promise. However, its complexity makes it difficult to trust, limiting its usefulness in both high-stakes and everyday applications.

This PhD project will focus on grounding common-sense planning and reasoning within rigorous, verifiable models. The goal is to develop novel formal methods to provide provable guarantees about agentic AI behaviour, moving beyond empirical testing towards a future of “correct-by-construction” AI systems. You will explore how to create formal specifications for complex tasks and develop techniques to ensure an agent’s actions are logically sound and aligned with its goals.

Candidate Profile: The ideal candidate will have a strong background in Mathematics, Computer Science or related discipline, with a deep intellectual curiosity for foundational research.

Essential: A first-class degree (or equivalent), a strong grasp of theoretical computer science and advanced mathematics, and excellent programming skills. A demonstrated, deep interest in learning and applying formal verification methods (e.g., model checking, theorem proving, type theory).

Desirable: Prior experience with formal methods (e.g., Coq, Lean), statistical learning theory, or AI safety research.

đź”” PhD Fellowship: Theoretical Foundations of World Models for Reinforcement Learning

Location: AI Centre, University College London (UCL) Start Date: September 2026

We are seeking a theoretically-minded PhD candidate for foundational research in model-based reinforcement learning (RL) at the UCL AI Centre. Current RL approaches face a critical challenge with data efficiency, often requiring complete retraining when environmental conditions or rewards change even slightly.

This project is built on the insight that world models provide a principled representation for enabling robust knowledge transfer and adaptive behaviour. By possessing a coherent, abstract model of its environment, an agent can generalize and adapt far more efficiently than one learning a purely reactive policy.

The core research will rigorously explore the mathematical foundations and operator-theoretic structures that govern how agents construct, reason with, and update internal world models. While motivated by practical challenges, the primary focus is on establishing a robust theoretical framework that transcends ad-hoc algorithms to foster a principled, model-based theory of agency. The successful candidate will validate their theoretical insights by implementing and testing these new constructs in single and multi-agent environments.

Candidate Profile: We are looking for a candidate with a strong aptitude for abstract reasoning and a desire to tackle the fundamental questions of AI.

Essential: A first-class degree (or equivalent) in Mathematics, Theoretical Computer Science, or a related field. A proven, strong background in advanced mathematics (e.g., abstract algebra, operator theory, probability). Excellent programming skills, with the ability to translate complex theoretical ideas into working code.

Desirable: Prior experience or coursework in reinforcement learning and game theory. Familiarity with modern deep learning frameworks.

How to Apply for PhD Fellowships

Important: These fellowships are open to candidates that have studied in the UK for 3+ years or have UK Nationality.

To apply for either position:

  1. Contact me directly at c.ciliberto@ucl.ac.uk with your CV and a brief expression of interest
  2. In parallel, submit an expression of interest through the UCL Computer Science PhD Programme

ELLIS PhD Programme

I am also affiliated with the ELLIS (European Laboratory for Learning and Intelligent Systems) network. Exceptional candidates may consider applying for an ELLIS PhD position, which offers unique opportunities for international collaboration and mobility within the ELLIS network of excellence. If you are interested in pursuing an ELLIS PhD under my supervision, please reach out to discuss potential projects.

General PhD Opportunities

I am also open to supervising PhD students on other topics related to machine learning theory, structured prediction, meta-learning, and their applications. If you have a strong research proposal that aligns with my interests, please feel free to reach out.

Postdoctoral Fellowships

Occasionally, I have open positions for postdocs. I also welcome ambitious postdoctoral researchers to join my group. Interested candidates to explore external fellowship opportunities and reach out to discuss potential collaborations or visits.

UK/International Fellowships:

EU Fellowships:

If you are interested in applying for any of these fellowships or would like to discuss other funding opportunities, please contact me. I am happy to support strong applications that align with my research interests.

Visiting PhD Students and Postdocs

I welcome visiting PhD students and postdocs for short-term research visits (3-6 months). Please contact me with your CV and research interests.