Luc McCutcheon

lucmccutcheon.home@gmail.com | GitHub | LinkedIn | Google Scholar

Research Engineer and PhD Candidate (completing Summer 2026) specialising in RL for distributed VLA post-training. Co-Founder/CTO of GeoSnap. Expert in JAX/PyTorch. Published at ICRA, NeurIPS, and IROS.

Experience

Research Scientist (Fixed-Term)

Cambridge Consultants — Oct 2025 – Mar 2026

  • Developed reinforcement learning pipelines for the Unitree G1 humanoid robot
  • Fine-tuned Vision-Language-Action models (GR00T) for manipulation tasks
  • Built simulation environments in MuJoCo and NVIDIA Isaac Sim
  • Designed and executed sim-to-real transfer experiments
  • Demonstrated VLA manipulation work at Mobile World Congress (MWC) 2026

Lead Research Scientist (Part-Time)

Agile Loop — Oct 2023 – Apr 2025

  • Led a team of 10 researchers across multiple agentic AI projects
  • Developed agentic pipelines combining RL and VLM LoRA fine-tuning
  • Presented technical results and product demos to Google Cloud stakeholders

Research Scientist (Part-Time)

Agile Loop — Jun 2023 – Oct 2023

  • Built RL environments for software testing and evaluation
  • Developed computer vision systems for product applications
  • Designed hybrid edge/cloud computer vision and reinforcement learning model routing architecture for Lenovo
  • Conducted in-context learning research using custom environments

Assistant Research Volunteer

Connected & Autonomous Vehicle Lab — Jun 2022 – Aug 2022

  • Implemented Noisy Neural Networks and Deep Q-Learning for autonomous driving
  • Tackled partial observability using LSTMs

Software Engineer Intern

QinetiQ — Jul 2018–Sep 2018 & Jul 2019–Sep 2019

  • Developed voice cryptography systems in C++ and Python
  • Conducted red team security assessments and penetration testing

Education

PhD Reinforcement Learning (Sponsored)

University of Surrey — 2021 – 2026 (expected)

Research focus: world models, Neural Lyapunov functions for stability, time-delay control systems, and large-scale JAX implementations. Published at top robotics and ML venues, supervised by S. Fallah.

BSc Computer Science (Hons) — First Class

University of Surrey — 2018 – 2021

CyberFirst Scholarship recipient. Strong foundation in algorithms, systems, and applied machine learning.

Skills

Frameworks & Libraries

JAX PyTorch Gymnasium vLLM NumPy

Languages

Python C++ Rust JavaScript

RL & ML

PPO GRPO LoRA VLA SAC RAINBOW ResNet

Infrastructure & Tools

Unitree G1 SLURM GCP AWS Docker

Languages (Human)

French (C1)

Publications

First Author

  • Neural Lyapunov Function Approximation with Self-Supervised Reinforcement Learning — ICRA 2025
  • Adaptive PD Control using Deep Reinforcement Learning for Local-Remote Teleoperation with Stochastic Time Delays — IROS 2023
  • Preventing Policy Collapse in Continual Reinforcement Learning — Under Review

Co-Author

  • Meta-World+: An Improved, Standardized, RL Benchmark — ICML 2025 Workshop (Spotlight) & NeurIPS 2025
  • In-Context Ensemble Learning from Pseudo Labels Improves Video-Language Models for Low-Level Workflow Understanding — NeurIPS 2024 Workshop
  • Prediction Based Decision Making for Autonomous Highway Driving — ITSC 2022

Reviewing

ICLR 2026 | NeurIPS 2025 | TNNLS 2025 | IROS 2025

Awards & Certifications

  • Honourable Mention — Grokathon 2026
  • Bronze Medal — Mathematics & Informatics Olympiad
  • Grace Hopper Award — University of Surrey
  • Foundership Award
  • NVIDIA Deep Learning Certificate
  • Coursera Reinforcement Learning Specialisation
  • Extended Project Qualification (EPQ)

Speaking

  • Google Agents Workshop — Guest speaker and panel discussion on agentic AI
  • University of Surrey — Model-based Reinforcement Learning; JAX vs PyTorch
  • Cambridge Consultants — Model Optimisation & Compilers; Safe Reinforcement Learning