Arnav Cherukuthota

- Developed and automated an end-to-end data pipeline to process financial data for December Tech, utilizing Python, SQL, and data from the NCUA (National Credit Union Administration) - CI/CD Automation: Built a GitHub Actions workflow to automate the deployment of infrastructure and the data pipeline, streamlining the development process - Serverless Pipeline: Deployed a data processing script as a serverless Cloud Run job, demonstrating the ability to build scalable and cost-effective batch processing solutions on GCP

- Formulating an event schedule for ColorStack through external outreach to potential partners, such as Humane Inc. - Campaigning ColorStack events by posting flyers and relaying in-class announcements, aimed at fostering a welcoming environment to provide technical education for Black and Latino/a computer science students

- Ideate and film creative reels to promote events hosted by the AI Student Collective (AISC) - Participate in weekly team standups to brainstorm a content plan to maximize outreach - Optimize student engagement and retention on AISC's discord server of over 500+ UC Davis students through advertising AISC's events, posting community announcements, as well as fostering AI-based discussion - Help manage and develop AISC's Instagram account of over 1,200+ followers

- Guided four researchers in building a fraud detection system using Random Forest, XGBoost, & Logistic Regression on a synthetic credit card transaction dataset - Conducted multiple forms of analysis, such as classification and regression, providing insight regarding credit card fraud hotspots, geospatial analysis, and socioeconomical feature importance - Contributed to a ten-page research paper explaining our team's project in-depth - Presented our work through an informative presentation during the AI Student Collective's Spring 2024 showcase

- Analyzed touchpoints across Calix's programs and platforms to find patterns for key customer segments regarding their educational content needs - Content Lifecycle Management: Created and updated Calix Customer Education content for specific personas, including the development of mobile microlearning modules, and working with content owners to sunset old material - Worked with a variety of internal departments to update key educational content and extend the reach of current materials into the Calix Community and beyond - Assisted with administration and optimization of the company's core learning management system

- Worked as an AI/ML developer on CourtCheck, a real-time tennis out-of-bounds detection app aimed at increasing fair play in collegiate and recreational tennis matches - Developed a Python algorithm to identify the location of the tennis ball, both players on the court, and prominent court point coordinates for thousands of images - TrackNet was used to detect the ball and extrapolate the ball's trajectory from image to image, clipping them together to create a seamless video of the ball's movement - Utilized a YOLOv8 object detection algorithm to detect the movement of the players from clip to clip, labeling either player as 'player 1' or 'player 2' - Implemented a Detectron2 algorithm for the detection of the court's corner points, aiding in ball out-of-bounds detection - Presented our project in the Aggie Sports Analytics Spring 2024 Case Competition, winning 2nd place while competing against 12 other teams

- Led a team of five student associates to build 'BeatCasso', an advanced song recommendation system based on user-specified criteria, such as valence, liveness, energy, acousticness, danceability, instrumentalness, & speechiness - Used pandas and scikit-learn to develop a K-Nearest Neighbors weighted classification algorithm - Utilized CSS, HTML, and Flask to construct a functional webpage allowing users to find song recommendations tailored to them - Engineered a scalable RESTful API and backend using Flask to produce real-time recommendations - Presented our project in two different showcases, one in March and one in June, winning the "Best Product Pitch" award as well as the "Most Innovative Project" award

- Worked in a team of nine as a student developer to build a tool for our client, MyEdMaster, that conducts semantic analysis on Amazon's "health and wellness" product listings - Webscraped data using BeautifulSoup and conducted correspondence analysis using Plotnine - Used Support Vector Classification (SVC) and tf-idf vectorization to predict sentiments for keywords

- Used sheets & and Excel to group, parse, and process customer data into nZero's stored database to propel sustainability and reach a net zero carbon footprint - Predictive Modeling: Used ggplot and dplyr (tidyverse) in RStudio to build visualizations that forecast future natural gas usage rates for certain clients based on prior usage data - Collaborated with product marketing and technology teams to demonstrate the app and the company’s goals to potential customers, including the Golden State Warriors - Wall & Roof Retrofitting: Worked under the data science team to program environmentally efficient wall and roof insulation retrofits using seaborn and CPLEX optimization in Python