ML lead for domain detections and revival monitoring

Employee number ~30. Acquired by Salesforce.

Lab: Intelligent Systems

Have been working on automating bug-fixing. Check out our paper here: https://arxiv.org/abs/2306.06490

• Conducted pioneering research to investigate the impact of edit-generation models on the accuracy and efficacy of code editing processes within the SARGAM framework

• Developed and evaluated different retrieval techniques, including TF-IDF, BM25, and PLBART embeddings, as baselines for patch retrieval, with each technique matched with similar modalities during retrieval

• Improved automated code editing performance through a sequential approach of generation followed by editing, as opposed to direct editing models

• Evaluated the proposed approach on two tasks: code editing and automated program repair, demonstrating its high effectiveness and outperformance of state-of-the-art methods in most settings