Huasen Wu learned how to develop a speech machine learning system around a real user need at UltraVoice in Philadelphia, Pennsylvania. That’s where Wu worked as a Machine Learning Engineer Intern. As a second year EE Master’s student with a concentration in Artificial Intelligence, Wu deeply valued how the internship gave him the ability to take ownership over a project.
“I worked across the full pipeline, including guided audio collection, data quality control, Whisper and LoRA model adaptation, evaluation, and deployment analysis,” Wu explains. “I also learned that carefully designed data and evaluation sets can be just as important as model architecture.” For Wu, the internship experience reinforced the significance of “accuracy, latency, reliability, and usability together when building an assistive speech product.”
Taking ownership of a project also gave Wu the ability to realize how his work “could directly improve communication for people who have undergone laryngectomy,” Wu emphasizes. “Working closely with the team and responding to real patient needs made the technical work especially meaningful.”
When Wu isn’t working, he enjoys working out ,exploring new technologies, spending time with friends, and playing table tennis.