Avery (Hee-Woon) Ryoo

education

MSc in Computer Science (AI Specialization), Mila / University of Montreal (Sep 2023 – Present)

• Supervised by Dr. Guillaume Lajoie and Dr. Matthew Perich

BASc in Biomedical Engineering (Computing Option), University of Waterloo (Sep 2018 – Apr 2023)

• Final-Year Design Project: Optimizing against cognitive fatigue in BCI-controlled wheelchairs

publications

• A. H. W. Ryoo, N. H. Krishna, X. Mao, M. G. Perich, G. Lajoie (2025). Towards generalizable, real-time decoders for brain-computer interfaces, COSYNE 2025 (Accepted Poster)
• E. Williams, A. H. W. Ryoo, T. Jiralerspong, M. G. Perich, L. Mazzucato, G. Lajoie (2024). Expressivity of Neural Networks with Random Weights and Learned Biases, ICML HiLD 2024, CCN 2024, ICLR 2025
• K. P. Patel, A. Ryoo, S. Mihalas, B. Tripp (2022). Topography of multisensory convergence throughout the mouse cortex, COSYNE 2023 (Under review for journal publication)

(* denotes equal contribution)

research experience

Graduate Research Assistant, Mila – Quebec AI Institute (Sep 2023 – Present)

• Main Project: Building multimodal deep learning approaches for neural decoding

Machine Learning Research Intern, Neurable (May 2022 – Aug 2022)

• Developed online-learning based denoising models for real-time EEG artifact removal
• Designed ML models to optimize BCI headphone signals

Computer Vision Scientist Intern, Hinge Health (Jan 2022 – Apr 2022)

• Developed temporal-based 2D human pose estimation models
• Improved accuracy by 0.08 AP with negligible computational overhead

other technical experience

AI Software Engineer Intern, WRNCH AI (May 2021 – Aug 2021)

• Developed a CoreML model surgery framework using coremltools and NumPy
• Designed a benchmarking pipeline for evaluating model output similarity pre- and post-conversion
• Optimized TensorFlow models for deployment on NVIDIA Jetson, Qualcomm Snapdragon, and iOS devices

Deep Learning Developer Intern, Applied Brain Research Inc. (Sep 2020 – Dec 2020)

• Developed classification and object detection models for windmill defect detection
• Implemented model optimizations using TensorRT quantization, reducing energy consumption by 300%
• Built a benchmarking suite for latency evaluation on various hardware configurations

Data Scientist Intern, Validere Technologies (Jan 2020 – Apr 2020)

• Created an automated anomaly detection tool for gas quality data, saving hours of manual calculations
• Developed a classification pipeline to process over 30,000 gas quality samples
• Provided data-driven recommendations that improved operational efficiency for gas plants

Junior R&D Engineer, Membio Inc. (May 2019 – Aug 2019)

• Led bioreactor design, 3D modeling, and experimental data acquisition
• Improved assembly process efficiency, reducing total build time by 54%

technical skills

Languages: Python, C++, C#, MATLAB
Frameworks: TensorFlow, Keras, NumPy, TensorRT, scikit-learn, OpenCV, ONNX
Tools: Git, Linux/Unix, Docker, Conda, Jupyter

awards & honours

• 2025: UNIQUE Excellence Scholarship – MSc ($10,000)
• 2022: Best Technical Project Award, Hinge Health Hack Day ($350)
• 2020: Undergraduate Research Assistantship, University of Waterloo ($600)

volunteering

• Workshop Organizer, COSYNE 2025: Organizer for “Building a foundation model for the brain”
• Organizer, UNIQUE Student Symposium 2024
• Volunteer, Montreal Neuro-AI Symposium 2023