Hi, I am Atharva Yeola.

Data Scientist

  • → A data scientist with expertise in analytics, decision science, machine learning, and LLMs.
  • Dr. John Ayers (his PI at UCSD) quotes:
  • → "Atharva is a Masters student at UCSD but easily mistaken for a senior PhD fellow.
  • → He is a wonderful addition to our innovation ecosystem at Qualcomm Institute (QI), bridging science and practice!
  • → In less than a year, Atharva published a study in JAMA Internal Medicine (4% acceptance rate), gaining significant media attention.
  • His work has been featured in several prominent media outlets, including:
    • New York Post
    • CBS News
    • NBC News
    • NPR
    • HIVTrends.org—a collaboration of CFAR, JHU, QI, and ACTRI—was made possible through his key contributions."

    Projects

    End2End Text Summarizer

    Built a text summarizer for concise summaries, managing the entire workflow from configuration to deployment. Automated CI/CD with AWS and GitHub Actions, handling Docker, EC2, ECR, and runner setup.

    • Transformers, NLP, GitHub Actions

    Retail Vision Enhancement

    Implemented YOLOv8 for object detection, achieving over 90% accuracy in labeling on-shelf retail products. Utilized SuperGlue for precise product identification, ensuring accurate differentiation between similar items. Additionally, a Dockerfile was created to guarantee cross-platform reproducibility, simplifying deployment and setup.

    • yoloV8
    • SuperGlue
    • Docker

    Skills

    • Python
    • SQL
    • Spark
    • AWS
    • Docker
    • LLMs
    • PyTorch

    News

    Original Investigation in JAMA Internal Medicine

    Our study titled "Growing Health Concern Regarding Gambling Addiction in the Age of Sportsbooks" was published in JAMA Internal Medicine. This research examines the association between the legalization of sports betting in the U.S. and the increase in gambling addiction help-seeking behaviors.

    February 17, 2025

    Launch of HIVTrends.org

    Collaborated on the development of HIVTrends.org, a platform providing public, real-time, and validated HIV testing sales trends from search query surveillance. This project aims to enhance public health monitoring and response strategies.

    December 20, 2024

    Research Study at IEEE Xplore (CVMI 2024)

    Our paper titled "Enhancing Traffic Sign Recognition: A Deep Learning Approach for Occluded Environments" was published in IEEE Xplore and presented at the 3rd IEEE CVMI 2024 conference.

    December 11, 2024

    Blogpost

    Optimizing Neural Networks: Pruning and Quantization Techniques

    This blog covers advanced techniques to optimize neural networks, focusing on pruning and quantization. Learn how these methods can reduce model complexity and improve performance.

    Read More

    August 10, 2024

    Contact

    Email me