Duration: May 2023 - Present
- Worked on Domain Generalization problem to extract domain invariant features for arrhythmia detection for 10 specific classes from hearts for 3 simulated ECG patient datasets.
- Implemented Domain Adversarial Neural Network (DANN) on simulated dataset with only 3 patients at one time to match the existing accuracy scores.
- Expanded research scope by applying the same architecture to 33 different clinical patients dataset, unveiling previously unexplored insights and achieving a notable 58% accuracy with expected validation loss curve.
Duration: May 2024 - Present
- Course Assistant for CSCI 431 - Foundations of Computer Vision which is an Undergraduate level course for Dr. Thomas Kinsman.
- Course Assistant for CSCI 631 - Foundations of Computer Vision which is a Graduate level course for Dr. Christian Yongwhan Lim.
- Course Assistant for CSCI 731 - Advanced Computer Vision which is a Graduate level courses for Dr. Thomas Kinsman.
- Provided detailed feedback to 70-80 undergrad and grad students to help them improve their understanding and application of Computer Vision concepts (Image processing and Video processing) for Foundations of CV and Advanced CV courses.
Duration: June 2020 - August 2022
- Designed and implemented end-to-end ML Automation pipelines using Scikit-Learn, Docker, GCP tools, and
Apache Airflow, advancing MLOps practices within the team.
- Constructed ETL pipelines and applied AI/ML models with statistical analysis using Python, Pandas, and Apache
Spark for large-scale data processing.
- Led the development of an Optical Character Recognition (OCR) system using Tesseract, PostgreSQL,
Docker, and SAP Cloud, by providing clients with Flask based RESTful APIs.
- Optimized GCP service utilization by strategically advising on cost-effective solutions, enhancing the department’s
cloud efficiency.