Machine Learning Intern - Harvard Biodesign Lab

Join our interdisciplinary team working at the intersection of machine learning, human movement, sensing, and rehabilitation. As a Research Intern in Machine Learning at the Harvard Biodesign Lab, you will develop and evaluate machine learning methods for analyzing complex human movement and sensor data. You will work with multimodal datasets collected from human participants and contribute to building models that can extract meaningful information about movement, function, and performance. Your work will involve developing data pipelines, training and evaluating models, designing experiments, and analyzing model performance, with the goal of advancing data-driven approaches for understanding human movement and informing future rehabilitation and assistive technologies. This internship is a 6-month training-focused, academic experience and is unpaid. Applicants typically join with their own fellowship, scholarship, university support, or personal funding. This internship must be completed for Master’s thesis credit (or an equivalent degree requirement), and candidates must be enrolled in a Master’s program for the duration of the internship. The position is based in Cambridge, MA and is in-person. **Description & Responsibilities** **Machine Learning Development:** Develop, train, and evaluate deep learning models for analyzing human movement, wearable sensor, and related multimodal datasets. **Data Processing:** Build and maintain pipelines for cleaning, preprocessing, structuring, and transforming datasets for machine learning experiments. **Model Evaluation:** Define appropriate evaluation metrics, benchmark model performance, perform error analysis, and investigate model strengths and failure modes. **Experimentation:** Design and run systematic machine learning experiments, including comparisons between model architectures, input representations, training strategies, and other methodological choices. **Reproducible Research:** Develop well-structured Python and PyTorch code and maintain clear experiment configurations, results, and documentation to enable reproducible analysis. **Research Communication:** Analyze and interpret experimental results and contribute figures, methods, and findings to internal reports, presentations, and research discussions. **Minimum Qualifications:** * Currently enrolled in a Master’s program in Computer Science, Machine Learning, Artificial Intelligence, Data Science, Electrical or Biomedical Engineering, or a related quantitative field. * Strong programming skills in Python. * Hands-on experience developing, training, and evaluating deep learning models through coursework, research, or independent projects. * Strong working knowledge of PyTorch, including implementing models, training loops, loss functions, optimization, and model evaluation. * Understanding of core machine learning and deep learning concepts, including training and validation procedures, overfitting, regularization, performance metrics, and model generalization. * Experience working with common scientific Python tools such as NumPy, pandas, scikit-learn, and Matplotlib. * Strong analytical skills, attention to detail, resourcefulness, and enthusiasm for interdisciplinary research. **Preferred Qualifications** * Experience working with time-series, sensor, biomechanical, computer vision, or other multimodal data. * Experience with modern deep learning architectures and representation learning methods. * Experience performing systematic model evaluation, hyperparameter tuning, ablation studies, or error analysis. * Experience training deep learning models on computing clusters or high-performance computing environments, including GPU-based workloads and batch scheduling systems such as Slurm. * Familiarity with Git, Linux, and tools for tracking and managing machine learning experiments. Previous research experience in machine learning, human movement, healthcare, robotics, or a related field. **Please apply through Sirop:** https://sirop.org/app/12dfc0e7-d9b9-4d3f-9afb-f9eb3503a0f2?\_k=d\_dl6cYbkS9t9muO