Edwin He

GPA: 3.76

Coursework: Machine Learning, Artificial Intelligence, Data Structures and Algorithms, Databases, Operating Systems, Computer Architecture, Python, C++, Java

• Developed automation scripts in Python to support regression testing and validate protocol performance for high-speed routing platforms
• Designed and maintained testbeds using traffic generators and protocol emulators, evaluating the fidelity of simulated scenarios against real-world network edge cases
• Investigated AI and LLM integration into Quality Assurance workflows to benchmark automated test generation and assist in log parsing for complex network scenarios

• Engineered a Mixture of Agents Large Language Model architecture utilizing a generate-critique-aggregate pattern to synthesize clinical insights from multimodal medical data, improving diagnostic reasoning accuracy
• Designed and deployed a cross-platform desktop framework for macOS, Windows, and Linux, integrating a secure cloud-based backend to manage remote license authentication, telemetry ingestion, and automated updates
• Architected a Retrieval-Augmented Generation (RAG) pipeline to ground model responses in verified clinical databases, reducing hallucinations by injecting real-time medical context into the inference prompt

• Fine-tuned transformer and classification models on multi-modal interaction datasets from study participants, achieving over 80% accuracy in categorizing student engagement and learning behaviors
• Processed, cleaned, and anonymized large-scale user interaction logs to track learning trajectories
• Synthesized and mapped relevant literature across Computer Science education and Human-Computer Interaction, establishing theoretical frameworks and baseline methodologies for experimental study design

• Annotated and reviewed large datasets for training and evaluating LLM programming capability, ensuring data accuracy and relevance
• Tested model outputs for correctness, consistency, and alignment with task instructions
• Provided detailed evidence-based feedback on model performance, including edge cases and anomalies

• Built a PyTorch-based training pipeline using 3D motion capture data and audio data, modeling root translation and joint rotation for audio-based full-body dance synthesis
• Converted 3D model hierarchy and motion channels into normalized tensor representations and aligned them with audio-derived features from audio files
• Trained temporal sequence models on synchronized audio–motion windows to generate beat-aligned dance motion