Incoming PhD · ETH Zurich · Visiting MIT

Luca Mouchel

AI researcher working on open LLMs, reasoning, and reinforcement learning. Worked on Apertus — Switzerland's open, compliant large language model.

CV Email
Portrait of Luca Mouchel
Experience

Where I've built

Full CV
Feb. 2026 – Aug. 2026 · Boston, MA

Visiting Researcher, MIT

Graph machine learning and AI labor exposure: modeling the US workforce by tasks and building retrieval-augmented pipelines over news and papers with open models served through vLLM.

  • Building task-level evidence systems for public-facing analysis of workforce change.
  • Work featured by NBC 10 Boston, the Boston Globe, and MIT CTL.
AI exposure RAG vLLM Labor analytics
July 2025 – Jan. 2026 · Boston, MA

Machine Learning Research Intern, SLB

Acoustics and complex systems team on Epilogue, SLB's well integrity product.

  • Integrated Vision Transformers into the production sonic log analysis pipeline; modular architecture trained from scratch with AutoKeras HPO.
  • Built GradCam++ interpretability workflows for CNN models on slowness-time coherence wavelet transforms.
  • Contributed to an ongoing patent application on the machine learning methods developed in this workstream.
Vision transformers TensorFlow AutoKeras Patent (ongoing)
2022 – 2025 · Lausanne, Switzerland

École Polytechnique Fédérale de Lausanne

Research across SwissAI, the AI Lab, and ML4ED.

Research Assistant — SwissAI Initiative

Jan. 2025 – July 2025

Post-training team for Apertus. Initialized online RL with VERL, built curriculum learning on Reasoning-Gym, scaled to 70B models on Ray, and contributed to QA/math evaluations with vLLM.

RL GRPO VERL LLMs

Research Scholar — AI Lab

Sept. 2023 – Feb. 2025

MSc Research Scholarship with Prof. Boi Faltings. Reasoning in LLMs, preference optimization for argument generation; co-authored work on epistemic consistency and uncertainty in causality.

Reasoning Alignment HuggingFace PyTorch

Research Assistant — ML4ED

Sept. 2022 – Feb. 2023

Bachelor thesis with Prof. Tanja Käser. Built a data mining and NLP pipeline to analyze self-regulated learning in keystroke logs.

Data analysis NLP EdTech
Education

Where I'm going next

Incoming

PhD, ETH Zurich

Starting 2026

Incoming doctoral student at ETH Zurich, continuing research in AI systems, reasoning, and large-scale model training.

Current

Visiting Student, MIT

Feb. 2026 – Aug. 2026

Visiting researcher in Boston working on AI labor exposure and retrieval-augmented research systems.

MSc Data Science, EPFL

Sept. 2023 – Aug. 2026

AI, NLP, reinforcement learning, and mathematics for data science. One of 15 MSc Research Scholarship recipients.

BSc Communication Systems, EPFL

Sept. 2020 – July 2023

Signal processing, stochastic models, ML, probability, algorithms, and software engineering. TA in Linear Algebra and Analysis III.

NAACL 2025 Outstanding Paper Award
EPFL MSc Research Scholarship
Service PC: AAAI, ACL, NeurIPS 2026
Selected Publications

Papers

Media

Coverage

NBC 10 Boston · June 23

Television segment on the MIT AI labor exposure project

Broadcast coverage of the research. Also featured by the Boston Globe and MIT Center for Transportation & Logistics.

Boston Globe article MIT CTL feature Map launch