CV

Basics

Name Pehuén Moure
Label Machine Learning and Neuroscience Researcher
Email pehuen@cornell.edu
Summary Engineer and scientist at the intersection of machine learning and neuroscience. Current work focuses on speech and accessibility; earlier work spans visual neuroprosthetics, adaptive speech recognition, and mechanistic understanding of large language models.

Work

  • 2026.09 - Present
    Postdoctoral Researcher, Runway Startup Postdoc
    Cornell Tech
    • Research on speech and accessibility for people with impaired speech
  • 2022.10 - 2026.06
    PhD Candidate in Information Technology and Electrical Engineering
    ETH Zurich - Institute of Neuroinformatics
    • Developed Bayesian LoRA fine-tuning for personalizing Whisper to speakers with speech impairments, with significant WER improvements on English dysarthric adults and a German child with structural impairment (ICASSP 2026, Interspeech 2026)
    • Showed where audio-language models fail to use multimodal context for dysarthric speech, and traced the failures mechanistically (COLM 2026, ICML 2026 workshop)
    • Co-developed a semantic re-chaining approach for adapting foundation ASR models to languages with little impaired-speech data (Interspeech DiSS Workshop 2025)
    • Proposed a Bayesian parameter uncertainty regularizer for RL agents with improved sim-to-real generalization (CVPR 2024)
    • Trained deep neural networks to control single-trial evoked responses in a blind participant's visual cortex, outperforming conventional prosthetic calibration at lower currents (Neuron 2026)
    • Designed a hybrid spiking autoencoder for temporally-varying cortical stimulation patterns in visual neuroprostheses (EMBC 2025)
    • Studied how extended reasoning aligns language model decisions with human probabilistic reasoning under uncertainty (NeurIPS 2026)
  • 2021.07 - 2021.10
    Research Scientist Intern
    Amazon Robotics
    • Developed RL agent for generative design of robotic warehouse layouts to support industrial engineer design
  • 2019.09 - 2020.10
    Data Scientist
    Amazon Robotics
    • Applied RL policy optimization to improve sortation center package throughput across multiple facilities
    • Developed adaptive sampling evolutionary algorithm to optimize sortation center floor maps for maximum efficiency
  • 2018.06 - 2019.09
    Software Development Engineer
    Amazon Robotics
    • Simulated new robotic solutions and quantified projected impact across fulfillment centers
    • Designed a data compute engine and data lake on Elastic MapReduce for large-scale operational analytics
  • 2017.08 - 2018.05
    Data Engineer Intern
    Amazon Robotics
  • 2017.06 - 2017.08
    Quantitative Finance Summer Analyst
    Morgan Stanley
    • Built an NLP chat-bot enabling traders to query and process data via natural language; implemented hierarchical clustering for automated trade recommendations

Education

  • 2022.10 - 2026.06
    PhD
    ETH Zurich - Institute of Neuroinformatics
    Electrical Engineering & Neuroscience
    • Elected graduate student representative at the Institute of Neuroinformatics
  • 2020.10 - 2022.05
    Master of Science
    University of California Los Angeles
    Electrical and Computer Engineering
    • Built an end-to-end system for computationally designing a fleet of unmanned underwater vehicles, coupling hull shape optimization with mission planning
    • Developed distributed RL for multi-agent path planning and coordination
  • 2014.09 - 2018.08
    Bachelor of Science
    Cornell University
    Computer Science
    • Meinig Family Cornell National Scholar; Morgan Stanley Robert B. Fisher Scholar
    • Built interfaces for socially assistive robots (facial tracking, audio localization); led 4-person undergrad team
    • Co-founded Suna Breakfast, a delivery startup for Cornell students; directed 10-person team building React Native apps and AWS backend
    • Co-founded Autonomous Bicycle Project; built SLAM-based object avoidance (ROS, TX-1, ZED stereo camera); managed 12 students across 4 sub-teams

Awards

Publications

2026

  1. Deep learning-based control of electrically evoked activity in human visual cortex
    Pehuen Moure*, Jacob Granley*, Fabrizio Grani*, Leili Soo, Antonio Lozano, and 6 more authors
    Neuron
    Neuron [2026] PDF Video
  2. When Audio-Language Models Fail to Leverage Multimodal Context for Dysarthric Speech Recognition
    Pehuen Moure*, Niclas Pokel*, Bilal Bounajma, Yingqiang Gao, Roman Böhringer, and 2 more authors
    In Conference on Language Modeling
    CoLM [2026] PDF Video
  3. Variational Low-Rank Adaptation for Personalized Impaired Speech Recognition
    Pehuen Moure*, Niclas Pokel*, Roman Böhringer, Shih-Chii Liu, and Yingqiang Gao
    In Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing
    ICASSP [2026] PDF Video
  4. Mechanistic Insights into Audio-Language Models for Impaired Speech
    Pehuen Moure, Bilal Bounajma, Niclas Pokel, Yingqiang Gao, Roman Böhringer, and 3 more authors
    In International Conference on Machine Learning Workshop on Machine Learning for Audio
    ICML [2026] PDF
  5. Localizing and Repairing Context-Failures in Audio-Language Models
    Pehuen Moure*, Niclas Pokel*, Bilal Bounajma, Longbiao Cheng, Yingqiang Gao, and 2 more authors
    In Conference on Language Modeling Workshop on Actionable Interpretability
    COLM [2026]
  6. Demonstration of Adapt4Me: An Uncertainty-Aware Authoring Environment for Personalizing Automatic Speech Recognition to Non-normative Speech
    Niclas Pokel, Yiming Zhao, Pehuen Moure, Yingqiang Gao, and Roman Böhringer
    In Proceedings of the ACM Conference on Conversational User Interfaces
    CUI [2026] PDF
  7. Data-Efficient ASR Personalization for Non-Normative Speech Using an Uncertainty-Based Phoneme Difficulty Score for Guided Sampling
    Niclas Pokel*, Roman Böhringer*, Yingqiang Gao*, and Pehuen Moure*
    In Proceedings of Interspeech
    Preprint at arXiv
    Interspeech [2026] PDF
  8. Reasoning Aligns Language Models to Human Cognition
    Gonçalo Guiomar*, Elia Torre*, Pehuen Moure, Victoria Shavina, Mario Giulianelli, and 2 more authors
    Preprint at arXiv
    [2026] PDF
  9. Active Probabilistic Reasoning in Humans and LLMs
    Gonçalo Guiomar, Elia Torre, Pehuen Moure, Victoria Shavina, Mario Giulianelli, and 2 more authors
    In Advances in Neural Information Processing Systems
    NeurIPS [2026]

2025

  1. Temporally-Varying Stimulations for Cortical Visual Neuroprosthetic using Spiking Neural Networks
    Pehuen Moure, Tatyana Pak, Niklas Hahn, Jasper Steveninck, and Shih-Chii Liu
    In International Conference of Engineering in Medicine and Biology Society
    EMBC [2025] PDF
  2. A Behavioral Study of Event-based Depth-Filtered Prosthetic Vision in Simulated Dynamic Environments
    Niklas Hahn, Pehuen Moure, and Shih-Chii Liu
    In International Conference of Engineering in Medicine and Biology Society
    EMBC [2025] PDF
  3. Adaptive Neural Control with Implicit Online Inference of Privileged Parameters
    Marcin Paluch, Florian Bolli, Pehuen Moure, Xiang Deng, and Tobi Delbruck
    In Learning for Dynamics and Control Conference
    L4DC [2025] PDF
  4. Adapting Foundation Speech Recognition Models to Impaired Speech: A Semantic Re-chaining Approach for Personalization of German Speech
    Niclas Pokel, Roman Böhringer*, Yingqiang Gao*, and Pehuen Moure*
    In Disfluency in Spontaneous Speech Workshop at Interspeech
    DISS [2025] PDF

2024

  1. Regularized Parameter Uncertainty for Improving Generalization in Reinforcement Learning
    Pehuen Moure, Longbiao Cheng, Joachim Ott, Zuowen Wang, and Shih-Chii Liu
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition
    CVPR [2024] PDF
  2. Bio-inspired parameter reuse: Exploiting inter-frame representation similarity with recurrence for accelerating temporal visual processing
    Zuowen Wang, Longbiao Cheng, Joachim Ott, Pehuen Moure, and Shih-Chii Liu
    In Proceedings of UniReps: the First Workshop on Unifying Representations in Neural Models
    NeurIPS [2024] PDF
  3. DeltaDEQ: Exploiting Heterogeneous Convergence for Accelerating Deep Equilibrium Iterations
    Zuowen Wang, Longbiao Cheng, Pehuen Moure, Niklas Hahn, and Shih-Chii Liu
    In Neural Information Processing Systems
    NeurIPS [2024] PDF

2021

  1. A Computational Design and Evaluation Tool for 3D Structures with Planar Surfaces
    Chang Liu, Wenzhong Yan, Pehuen Moure, Cody Fan, and Ankur Mehta
    In International Conference on Robotics and Automation
    ICRA [2021] PDF

2018

  1. Can Interactive Systems Be Designed for Conviviality? A Case Study
    Marc Choueiri, Schuyler Duffy, Sanjay Guria, Conrad McCarthy, Pehuen Moure, and 4 more authors
    In Proceedings of the 2018 ACM Conference Companion Publication on Designing Interactive Systems
    DIS [2018]

Volunteer

  • 2019.09 - 2020.08
    Pediatric Oncology Research Assistant
    Dana Farber Cancer Institute
    • Developed pipeline for automating analysis of chip-seq and RNA-seq data
    • Interacted with team of scientists to build tools that speed up production of results

Skills

Machine Learning
Reinforcement Learning
Deep Neural Networks
Bayesian Neural Networks
Computer Vision
Software Engineering
Python
AWS
Agile Development
ROS

Languages

English
Fluent
Spanish
Native

Projects

  • 2024.01 - 2024.06
    Speech Recognition for Children with Speech Impairments
    Developed tool for improving speech recognition accuracy in children with speech impairments
    • Implemented Whisper model with bayesian mechanism to improve accuracy on small datasets
    • Coordinated team of 5 students to develop tool for real-time speech recognition
  • 2016.07 - 2018.08
    Autonomous Bicycle Project
    Co-founded and led software team for autonomous bicycle development
    • Built object avoidance system using SLAM algorithm in ROS with TX-1 and ZED stereo camera
    • Managed team of 12 students working to create centralized code structure across 4 sub-teams