Madhura Deshpande

ML @ CrowdStrike | AI @ CMU

Pittsburgh, Pennsylvania, United States

About

I am a graduate student at Carnegie Mellon University specializing in Artificial Intelligence.

Experience

  • Machine Learning Engineer at CrowdStrike
    Jun 2025 - Present · 1 yr 1 mo

  • Carnegie Mellon University School of Computer Science (9 mos)
    • Graduate Teaching Assistant
      Jan 2025 - May 2025 · 5 mos

      TA for 10/423 - Generative AI (Machine Learning Department) As a Teaching Assistant under Prof. Matt Gormley and Prof. Pat Virtue, I am responsible for designing assignments and coding tasks, grading student submissions, and providing project mentorship.

    • Graduate Teaching Assistant
      Sep 2024 - Dec 2024 · 4 mos

      TA for 11-777 Multimodal Machine Learning As a Teaching Assistant under the guidance of Prof. Daniel Fried and Prof. Yonatan Bisk, my responsibilities include mentoring multiple student teams, guiding them through their projects, providing feedback, and grading their assignments.

  • Machine Learning Engineer at LUCA.ai
    Sep 2024 - May 2025 · 9 mos

    Conversational AI Tutor | Capstone Project at LUCA.ai Designing an interactive, speech-driven AI reading companion to support K–12 students with dyslexia by improving reading fluency, phonemic awareness, and learner engagement.

  • Machine Learning Engineer Intern at CrowdStrike
    May 2024 - Aug 2024 · 4 mos

    • Created and launched a Slack Bot that enhanced internal workflows with two core features:  – Automated Jira ticket creation, streamlining processes and reducing manual effort  – Real-time query resolution via integration with CrowdStrike’s custom LLM, saving employees ~10 hours/week • Developed an advanced end-to-end RAG pipeline using internal databases to significantly improve LLM response accuracy by 35% based on user feedback • Recognized with the "Relentless Focus on Innovation" award for outstanding contributions, selected from a cohort of 20 interns

  • Tel Aviv University (Tel Aviv University, Israel)
    • Machine Learning Researcher
      Mar 2023 - Aug 2023 · 6 mos

      • Worked under the guidance of Prof. Irad Ben-Gal (TAU) and Prof. Nicholas Bambos (Stanford) as part of the Digital Living 2030: TAU–Stanford Partnership initiative • Led a peer-to-peer federated learning research project that removed the need for a central server, focusing on distributed optimization across clients • Investigated the impact of data-sharing schemes, network topologies, and communication strategies in both IID and non-IID client settings • Proposed novel inter-client data-sharing mechanisms that accelerated model convergence and improved performance across varying data complexity levels

    • NLP Researcher
      Sep 2022 - Mar 2023 · 7 mos

      • Worked as a researcher at the TAD Center for Artificial Intelligence & Data Science, under the mentorship of Prof. Roded Sharan, Prof. Ronen Avraham, and Prof. Tamar Kricheli-Katz, on a collaborative project between the Buchmann Faculty of Law and the Blavatnik School of Computer Science at Tel Aviv University. • Established an automated data pipeline using PyTorch for preprocessing legal textual data - United States Supreme Court (U.S.S.C) opinions. • Modeled BERT-based NLP classifier for Judicial Authorship Identification with an accuracy of 76%. • Contributed significantly to securing the ‘TAD Center 2023 Research Grant’ advancing innovative research and fostering interdisciplinary collaboration.

    • Undergraduate Student Researcher
      Mar 2021 - Aug 2022 · 1 yr 6 mos

      • Conducted research at the Laboratory for AI, Machine Learning, Business, and Data Analytics (LAMBDA) under the guidance of Prof. Irad-Ben Gal and Prof. Parteek Bhatia. This research also contributed to my Final Year Thesis at Tel Aviv University. • Developed a computer-vision-based system to inform educators about learners’ attentiveness and understanding levels in an online education setting—under the guidance of Prof. Irad Ben-Gal (Head, LAMBDA). • Implemented deep network architectures improving on VGG16, ResNet-50, and Convolutional Neural Networks using TensorFlow and Keras in Python to predict learners’ engagement and affective states. • Designed and built a novel mathematical model employing ML algorithms to determine a comprehensive attentiveness metric for learners. • Conducted the full stack development and deployment of the neural network-based dynamic and user-friendly web API on a cloud server.