Harshika Jalan Email & Phone Number

Berkeley, California, United States

• Conducted interdisciplinary research at the intersection of robotics, AI, and healthcare, focusing on automating sutureplanning for robotic surgery.• Spring 2023: Developed a novel algorithm to optimize surgical suture placement for two-dimensional wounds. Evaluatedits performance on wounds of varying degrees and curvatures.• Fall 2023/Spring 2024: Extended the algorithm for three-dimensional applications by leveraging deep learning modelsfor wound segmentation and generating 3D meshes. Modified optimization constraints and objectives for 3D space.• Won Best Healthcare Automation Award at IEEE CASE 2024 and the Best Student Paper Award at IEEE CASE 2023.

San Francisco, California, United States

• Designed and built a scala service for Envoy to send requests to the Ratelimiting Service.• Added API Tests and load tests to exhaustively test the integration and determine memory and CPU requirements.

Santa Monica, California, United States

• Created and scaled data pipeline using Apache Beam, Google Big Query and Airflow to help train ML models.• Served content moderation machine learning model into production and analysed model performance on Snapchat stories.• Updated the recommendation backend code to use model signals for ranking and feed diversification.

Berkeley, California, United States

• Worked in a team to develop keyword extraction algorithms that identify sub-categories among companies in a given industry.• Generated semantic word embeddings for companies using the BERT language model, along with K-means algorithm to cluster together similar companies.• Experimented with keyword generating algorithms such as TF-IDF, TextRank and RAKE.

Berkeley, California, United States

Lab assistant for CS61A, the introductory CS course, for the Fall 2021 semester. Furthered student’s knowledge of topics such as higher order functions and recursion in labs and office hours.

Singapore

• Developed a system using deep learning and computer vision techniques to automate the process of checking surgical instruments and reduce the number of incomplete surgical sets in operation theaters.• Trained an image identification model on a surgical tools dataset and incorporated the model into a web application built using HTML, JavaScript, Python and Flask.• Authored a technical research report and presented to a team of nurses at the Singapore General Hospital