Innovative Machine Learning Developer with substantial experience in designing, implementing, and improving machine learning models. Skilled at solving complex problems using advanced statistical techniques and programming tools. Strengths include a deep understanding of data structures, algorithms, computational complexity, and software development methodologies. Previous work has significantly contributed to optimizing predictive modeling processes and enhancing system performance.
Deep Learning Team Leader & Principal Machine Learning Developer May 2021 – Now
§ Lead the research & development for the Deep Learning team.
§ Work with the upper management, internal and outside customers, and collaborate with other teams to establish goals for the team, prioritize tasks and assign tasks to developers.
§ Provide professional and moral support for the team by sharing knowledge and experiences, providing guidance, removing major obstacles, offering praise and feedback, providing opportunities to showcase their work, etc.
§ Run the scrum meetings, monitor the development process, adjust stories when needed, provide feedback for issues, and report progress to the upper management.
§ Collaborate with the testing team, product/project management to ensure smooth delivery of products in the CI/CD pipeline.
§ Lead the team to the successful delivery of the first release of Pytorch based deep learning action set (dlModelzoo) in Sept. 2022 according to schedule.
§ Continuously improve and add more features, as well as DL models to the dlModelzoo action set through a CI/CD pipeline.
§ Developed a hybrid multi-process-based framework for hype parameter tuning and training for deep learning models.
§ Developed a mechanism to incorporate Torchscript based deep learning models in the DlModelzoo action set.
§ Developed a mechanism to incorporate and customize Python Pytorch models in the DlModelzoo action set.
§ Developing patents for innovative research and development.
While working towards my Ph.D., I worked to convert my research in computer vision/biomedical image processing into commercial software for the start-up company.
Helped to deliver an innovative video-based eye-tracking system for medical applications with state-of-the-art performance and increase its market share.
§ Developed a robust high-speed eye-tracking system with a higher frame rate (200 Hz x 2) and higher accuracy than available online video-based eye-tracking systems using only off-the-shelf computer system and digital cameras.
§ Designed a robust torsion eye-tracking system with high performance in terms of frame rate and accuracy.
§ Developed a novel computer vision calibration system with better accuracy by utilizing the pupil shape at different eye position.
§ Developed a pupil size monitoring and feature extraction system for the screening of Diabetic Retinopathy (DR).