Used advanced computer vision techniques and frameworks (PyTorch) to find a scale-invariant, Markovian state space representation of cooking. Synthesized a massive dataset with thousands of multimodal data points augmented with auto-generated labels via LLMs, trained a state-of-the-art visual classifier for food using transfer learning and contrastive learning techniques. Developed GPU and RAM optimized code to train model given limited compute; adapted real-world data (blurry and disorganized unlabeled video footage, incomprehensible machine-generated cooking logs, etc.) to structured formats using data augmentation and preprocessing techniques. Developed novel model architectures incorporating CNN and transformer-based components. Discussed ideas with CloudChef engineers (top IIT graduates) and implemented the solutions independently.
Research Assistant
Yale Vision Laboratory
Worked as research assistant at Yale (Computer) Vision Laboratory under Prof. Alex Wong; was involved in a PyTorch project involving tracking a robot performing an anastomosis surgery using convolutional neural architectures.
Analyst Intern
Appian Corporation
Interned at Appian (a multibillion-dollar public cloud-computing and enterprise software company) directly for founder Marc Wilson. Conducted interviews and leveraged tools such as Salesforce to develop a data-driven executive engagement program for the company. Participated in company-critical meetings at Appian and performed financial analysis on key Appian accounts. Designed a program that is now fully implemented across the entire 2,500-employee company and has led to a new Office of Executive Engagement. Received offer to return to employment at Appian.