Apoorva Gondimalla

- Hardware and Wearable Tech - Health Sensing & ML - Research Ops

Facilitating evaluation of the effectiveness of LLM-based fact-checking reports in detecting misinformation, through a survey study.

Led two research initiatives aimed at designing technology solutions and improving operational efficiency for case workers at the City of Austin's homelessness case management service. - Collaborated with data scientists to develop Large Language Model (LLM)-based generative summaries that are "useful" to the caseworkers, through co-design sessions, interviews, interface, and prompt design. - Identified design implications in building technology solutions for data collection by caseworkers through storyboards, speed dating, and in-depth interviews with multiple stakeholders. - Uncovered current note-taking patterns using advanced statistical analysis and qualitative coding of 50,000 unstructured clinical case notes, informing design opportunities. - Curated an executive report providing insights and recommendations on organizational data practices.

Compilers CS 375 - Managed grades of 50+ students in an accelerated summer course and held consulting hours.

- Generated insights on the usability and utility of a search tool for idea generation by managing and moderating over 60 user studies. - Assisted in refining the experimental study design through pilot studies and continuous feedback. - Best paper CHIIR 23

- Evaluated user experience of Autonomous Vehicle fleet operation tools at Motional utilizing qualitative and quantitative research methods. - Identified usability and scalability issues, and growth opportunities, driving 5 initiatives into the quarterly product roadmap. - Enabled utilizing quantitative user data for product decisions by establishing UX metrics, conducting workshops, and cross-functional collaboration. - Created dashboards, documentation, presentations, and UX artifacts.

- Increased click-conversion for the new feature by 40% by providing design recommendations through user studies, and analyzing post-launch user behavior using Google Analytics and Mixpanel data. - Crafted an interactive Tableau dashboard highlighting user engagement for strategic decision-making and KPI reporting.

-Improved text recognition from retail invoices by communicating insights from exploratory data analysis, building a synthetic data generator in Python to accelerate training, and systematic evaluation of encoder-decoder models, in an agile environment. - Resulted in a live product serving 40+ retail customers.

Improved performance of deep learning models for video classification by evaluating performance data, accuracy measures, hyperparameter tuning, and benchmarking.

Developed an action recognition model with 80% accuracy, enhancing monitoring of automobile manufacturing activities. Pre-processed data, trained CNNs, evaluated outputs, and utilized data augmentation techniques on a small and skewed dataset.

Designed a natural interface game, through field studies, ideation, and user testing, addressing the accessibility challenges for stroke victims in handling digital devices for rehabilitation games.

Redesigned the website for the visitor gate pass system, resulting in faster and more efficient visitor logging and tracking.

Thesis: Aligning Data with the Goals of an Organization and Its Workers: Designing Data Labeling for Social Service Case Notes. Courses: Human-AI Interaction, Usability, User Behavior and Search Experience, Data Visualization, Data Storytelling, Information Architecture, Interaction Design, Inclusive Design, Ethnography.

Thesis: Human action recognition from RGB videos using Convolutional Neural Networks. Relevant courses: Data Mining, Information Retrieval, Database Management Systems, Probability and Statistics. Activities: Core member at BITS Alumni Affairs Division(BITSAAD). Member of the Environmental Protection and Conservation Club(EPAC)