Diya Kini

Incoming Summer 2026 Intern

Incoming Fall 2026 Intern

- Integrated computer vision into an automated liquid dispenser, achieving 95% accuracy in hand gesture recognition using MediaPipe Hands for real-time control of dispensing, allowing users to select liquid portions with hand signals - Developed a finger tracking system that reliably detected 1-3 fingers and triggered liquid dispensing, improving user interaction by enabling hands-free operation with 86% reliability in various lighting conditions

- Applied spatiotemporal deep learning models (STGCN, GAT, TGN) in Python to analyze aircraft trajectory data, conducting experiments with labeling, Docker and Podman containerization, CI/CD, and agile development - Created an AI assistant to automate the process of software engineers retrieving UCI message entity translations for internal customers through Python scripting, web scraping, Genesis, RAG, and Gitlab API, saving ~$27,000 per team member by eliminating request processing and boosting engineer efficiency by 300% with a 1:1 replacement for query processes - Developed a Python WebSocket communication layer to interface with a C++ Open Mission Systems (OMS) server, enabling CAL compatibility and facilitating seamless integration with next-generation avionics platforms

- Executed matrix and vector operations and linear transformations to optimize LiDAR data interpretations in robotic navigation tasks - Taught 200 students the application of ML concepts, including linear regression, to develop predictive models for robotic systems, enhancing pattern recognition and decision-making processes - Explained feedback control algorithms for dynamical systems, such as balancing a Segway robot, leveraging linear algebra and Julia programming for improved system stability and performance

- Specialized in improving the Stereo vision for the Jetson Nano through ORB SLAM of the MBot ecosystem for hands-on, graduate-level robotics courses - Led the development of a vision-based robot for autonomous movement with OpenCV and Python

- Analyzed the optimal length of time a novel gene therapy would keep hemophilia patients at a safe level of clotting through the R programming language - Worked with a team of 3 data scientists to investigate the relationship between biomarkers of the gut biome and clinical outcomes for an innovative treatment for C-difficile infections - Assisted in the FDA approval of 2 treatments by developing plots of drug efficiency and safety alongside PhD Statisticians