Served as the Associate Academic Director for MScAC by reviewing projects and grant proposals, student study plans and final reports, and liaising between faculties members across departments.Led the research and academic excellence of the MScAC program by strategically minimising risk and maximizing quality by refining and executing the admissions filtering process, developing professional skills and internship/career coaching.Notable successes included helping to nearly double the program size, introducing the AI concentration (2022), the Data Science for Biology concentration (2023), and the AI for Healthcare concentration (2023).

Implemented transfer learning for adversarial neural architecture by strategically research latest advancements to integrate with team expertise, adapting models from state-of-the-art, and conducting model optimizing and parameter tuning. Produced the state-of-the-art results for transfer learning by combining multi-task meta-learning techniques with neural interpretability of neural language model thereby advancing the boundary for text classification.

Directed data initiatives by mapping out overall process, establishing project milestones, delivering baseline and prototypes within time line, and iterating upon sub-components.Established team cohesiveness by spearheading daily team scrum for member transparency, participating in cross-functional meeting, and communicating research advances to senior executive to influence direction.Oversaw research collaborations by establishing academic relationship to solve business problems, writing and applying for grant funding, coordinating legal approval process, and managing project progress such as coordinating data transfer and troubleshooting issues.

Led the sentiment analysis project end-to-end by starting with exploratory analysis, selecting appropriate algorithmic techniques, evaluating model, and productionized pipeline to produce generated data reports and presentations for internal stakeholders and external industry conferences.Drove the collection of high-quality training dataset by seeking internal expert annotators for pilot annotations and using crowd-based annotators, while maintaining external worker quality and satisfaction, thus producing supervised models that are reproducible and measurable.