Arjun Ravi Shankar

Led cross-functional teams in developing computer vision models for surgical video from ideation to production. Collaborated with team of deep learning engineers to create pipelines for expedited model development. - Enabled rapid, deep learning model development and 30% reduction in data labeling costs by designing and implementing fully automated cloud-based active learning pipeline in conjunction with team of 4 data scientists. - Directed 10-person team of clinicians, clinical engineers and annotators, building a state-of-the-art scalable production-ready computer vision model for detecting bodily fluids in surgical video with accuracy and F1 of 90%.Led cross-functional teams in developing computer vision models for surgical video from ideation to production. Collaborated with team of deep learning engineers to create pipelines for expedited model development. - Enabled rapid, deep learning model development and 30% reduction in data labeling costs by designing and implementing fully automated cloud-based active learning pipeline in conjunction with team of 4 data scientists. - Directed 10-person team of clinicians, clinical engineers and annotators, building a state-of-the-art scalable production-ready computer vision model for detecting bodily fluids in surgical video with accuracy and F1 of 90%. Skills:

Constructed multimodal ML models for lung cancer nodule classification and managed relationships with external AI partners. Trained in leadership skills as part of selective R&D Leadership Development Program. - Shortlisted potential lung cancer deep learning partners by orchestrating rigorous predictive modeling competition, resulting in strategic collaboration between Optellum and J&J. - Created multimodal Elastic Net and SVM models, classifying lung nodule malignancy using CT imaging features and blood-based miRNA expression levels. - Implemented LSTM model, predicting future risk of lung cancer from past EMR data.

Collaborated with team of vaccine shop floor engineers, director of operations and quality experts, collecting user stories and characterizing vaccine alarm management problem. -Formulated strategies for time series forecasting (LSTM), predicting vaccine batch efficacy.