Ryan Pégoud

PhD Candidate (BMW ProMotion)

Professional Summary

I’m currently pursuing a PhD as part of the BMW ProMotion program under the supervision of Dr. Nael Fasfous (BMW) and Prof. Dr. Daniel Müller-Gritschneder (TU Wien). My research focuses on the efficient deployment of end-to-end autonomous driving models, particularly on the compression of vision-language models and lightweight test-time adaptation.

Previously, I graduated from UCL’s computational statistics and machine learning MSc with and distinction and completed my thesis under the supervision of Prof. Tim Rocktäschel. I also hold an MEng in computer engineering from EPF Engineering School, where I later intervened as a lecturer and module leader for the natural language processing course. During my MEng, I completed two internships as a data scientist at BMW and CEWE.

Download CV

Education

PhD (BMW ProMotion Program)

TU Wien

MSc Computational Statistics and Machine Learning

University College London

MEng Computer Engineering

EPF Engineering School

Interests

Autonomous Driving Natural Language Processing Reinforcement Learning GPU Programming

Experience

Module Leader, Lecturer and Pedagogical Tutor
EPF Engineering School September 2023 – June 2024
- As Lecturer (09/2023 – 01/2024): Designed and delivered a new MSc-level NLP module for 30 students, creating the curriculum from scratch and managing all lectures, assignments, and exams.
- As Tutor (01/2024 – 06/2024): Mentored an MSc student through their thesis on Retrieval-Augmented Generation (RAG) for helicopter certification, in partnership with Airbus Helicopter.
Data Scientist Intern
BMW Group February 2023 – August 2023
Development of an anomaly detection system based on time-series data from BMW cars. Deployed a real-time, explainable and quantifiable solution at the scale of the BMW fleet with PySpark and AWS, replacing a low-performing LightGBM model.
Data Scientist Intern
CEWE July 2021 – January 2022
Development of a multilingual BERT model for text classification achieving a 90% F1-score across 14 classes and 3 languages. Built an active learning platform with Dash and implemented temperature scaling to ensure the model’s long-term sustainability and calibration.
Data Scientist Intern
CEWE June 2022 – July 2022
Shipped two projects: an aspect-based sentiment analysis pipeline with BERT and a time series model using NeuralProphet to forecast customer service traffic.

Education

PhD (BMW ProMotion Program)
TU Wien February 2026 – February 2029
Research in efficient deployment of end-to-end autonomous driving models.
MSc Computational Statistics and Machine Learning
University College London September 2024 – September 2025
Thesis on Memory Augmentation for Agentic Language Models. Supervised by Prof. Tim Rocktäschel. Yearly Average: 78.5% (Distinction)
Read Thesis
MEng Computer Engineering
EPF Engineering School March 2019 – June 2024
Thesis on Time Series based anomaly detection for fleet connectivity at BMW Group. Yearly Average: 84.6% (Distinction)

Conference Papers and Workshop Publications

Ryan Sullivan, Ryan Pégoud, Ameen Ur Rahmen, Xinchen Yang, Junyun Huang, Aayush Verma, Nistha Mitra, John P. Dickerson (2025). Syllabus: Portable Curricula for Reinforcement Learning Agents (Outstanding Paper Award). In RLC.

PDF Code DOI

Ryan Pégoud, Thibault Lahire (2024). Better Gradient Steps for Deep On-Policy Reinforcement Learning. In ARLET workshop at ICML.

PDF

Online Publications with Editorial Review