• Interned at SacHacks, the biggest intercollegiate hackathon in Sacramento to develop programs that assists them in handling logistics• Revamped Gavel, a hackathon project judging system by rebuilding the backend using Flask, PostgreSQL and Python contributing to a 15% increase in data retrieval and storage speed• Utilized HTML, CSS and Sass to develop a user-friendly frontend which streamlined the judging process by 25%• Deployed the webapp through Vagrant and Oracle VirtualBox which reduced the loading time of the application by 20%

- Analyzed the internal engine of ClabAI and constructed detailed developer and user flows using Notion- Compiled 2 informational college PDFs containing 29,000 words into a GPT-4o driven Microsoft GraphRAG model, achieving an output of 9.91K+ Tokens Per Minute and 240 API calls to the OpenAI API- Integrated Docker, OpenSearch and Supabase to query 2635+ records of college application tips. Fed the data into 5 Reranking algorithms from LlamaIndex and Cohere and achieved an average retrieval time of 7.1 seconds- Utilized Pandas, Jupyter Notebook and Poetry to finetune a custom Named Entity Recognition model from GliNER using 100 labeled sentences as the training dataset

• Interning at Aggie Sports Analytics, the first sports analytics club at UC Davis with the objective of leading a team to develop an application for a well established client• Spearheading Garde, a full-stack web app (MERN stack) utilized by Davis Fencing Academy, a Team USA fencing certified institution to improve fencer form by 35%• Engineering a robust backend with Node.JS and MongoDB to accelerate data processing by 30% and creating an immersive frontend using React to increase fencer’s engagement by 20%• Utilizing JIRA Kanban Boards to assign 20+ stories to facilitate agile development methodologies and enhance productivity and collaboration of the developer team by 65%• Leveraging Tensorflow.JS, OpenCV.JS, Mediapipe and OpenAI API to amplify the precision of measuring a fencer’s posture and expedite the feedback process by 15%• Mentored 5+ attendees in a React personal portfolio workshop.

• Developed and implemented a cutting-edge object detection model using the YOLO algorithm and a dataset of 800+ images, enabling real-time basketball detection in videos.• Employed PoseNet, a deep learning TensorFlow model, to accurately label the head, elbows, and knees of basketball players in real-time, effectively optimizing their shot form during the release of the basketball.

• Interned at Obbserv, one of India’s best digital marketing agencies to create a program which promotes social awareness• Demonstrated proficiency in algorithm design and analysis, leading to my findings being used in the ongoing initiative, “Clean-Jal” – an initiative dedicated to combatting plastic waste in water bodies• Produced comprehensive flowcharts and wireframes illustrating strategies for effective retrieval of hard plastic waste from the lake, enabling authorities to increase plastic retrieval time by 25%• Leveraged EmguCV and utilized 5+ image processing filters, resulting in an 85% boost of the accurate identification of hard plastic waste in Lake Pichola

Koursie was the first tech-based startup I worked in, although we weren't able to propel it into greater heights, I still was able to leave my mark by creating an email verification system that uses MongoDB and C#.