Dayne Lee
Jersey City,USA
Overview
5
5
years of professional experience
Work History
Research Intern
NYU Langone Health
New York, NY, USA
06.2025 - Current
- Developed multimodal machine learning model on 8K+ brain MRIs (T1, T2, FLAIR, DWI) and clinical data to estimate brain age, supporting early detection of neurological disorders.
- Optimized Docker builds (↓ 87% image size, ↓ 98% build time) and integrated automated end-to-end testing in CI/CD for open-source MRI preprocessing software, improving reproducibility and driving 3x more GitHub activity.
Machine Learning Engineer
Return Zero Inc. (Series B speech-to-text startup)
Seoul, South Korea
12.2021 - 08.2022
- Led end-to-end ML development for sound event detection, earning top-4 finish in leading international benchmark in audio recognition (DCASE Challenge).
- Cut infrastructure cost by reducing network traffic 75% in CNN-based voice activity detection (VAD) system, optimizing processing on Triton Inference Server.
- Improved evaluation performance (↓ 13 pp on DCF, ↓ 4 pp on FPR for music) for VAD, designing 1-model/2-worker setup and integrating cross-domain data (AMI meeting recordings).
- Enhanced event tracking and reliability for VAD by developing data preprocessing pipeline and automating evaluation.
Software Engineer
Return Zero Inc. (Series B speech-to-text startup)
Seoul, South Korea
08.2020 - 12.2021
- Developed and deployed in-house annotation platform (Node.js, Vue.js), cutting vendor costs, boosting labeling efficiency, and producing 10K+ hrs high-quality data; partnered with research team on feature delivery.
- Reduced query latency and enabled faster product iterations, redesigning schema and migrating MySQL to MongoDB.
- Cut troubleshooting time 50% by building logging system and implementing concurrency controls to resolve recurring crashes.
- Collaborated with PM, design, and backend in Agile sprints to ship B2B speech recognition pilot using React.
- Reduced annotation errors by 5%, implementing custom job scheduler in Go to automate invalid label detection.
Education
M.S. - Data Science
New York University
New York, NY09.2024
B.S. - Computer Science & Engineering
Kyungpook National University
08.2023
Skills
- Languages: Python, SQL, JavaScript, C/C, HTML/CSS
- Technologies: Docker, Linux, Conda, MySQL, MongoDB, Git, CircleCI, Spark, Google Cloud Platform (GCP)
- Libraries: PyTorch, TensorFlow, scikit-learn, SciPy, NumPy, Pandas, pytest, Matplotlib, Expressjs, Nextjs, Vuejs, React
Timeline
Research Intern
NYU Langone Health
06.2025 - Current
Machine Learning Engineer
Return Zero Inc. (Series B speech-to-text startup)
12.2021 - 08.2022
Software Engineer
Return Zero Inc. (Series B speech-to-text startup)
08.2020 - 12.2021
B.S. - Computer Science & Engineering
Kyungpook National University
M.S. - Data Science
New York University