Ethan Chandra

Experience

• 

  NASA - National Aeronautics and Space Administration

  May 2021 - Aug 2021

  NASA SEES Research Intern

  • Conducted 150 hours of research in completion of Gravity Recovery and Climate Experiment (GRACE) data analysis project• Formulated final project to evaluate the effectiveness of convolutional neural network architecture using different input data representations and publicized results live in NASA’s Virtual Science Symposium as part of the GRACE team with 5 interns• Oversaw technical discussions over program development and resolved programming problems team members encountered

• 

  Texas Guadaloop

  Jan 2023 - now

  Research Executive

  • Spearheaded a research team to contribute to U.S. hyperloop sustainability development, becoming one of the few U.S. university teams focused on novel hyperloop feasibility research for the Texas Triangle region• Competed in European Hyperloop Week (EHW) 2024 at Zürich and placed as a finalist for the Full-Scale: Socioeconomic category (top 5 out of 12 competing teams); presented and discussed our research on “Hyperloop Sustainability in the Texas Triangle”• Implemented Agile methodologies in Monday.com to streamline research processes across environmental, financial, and economic project teams, resulting in a 40% increase in research deliverables accomplished and project delivery speed Show less

• 

  Klein Training App

  Oct 2024 - now

  AI Research Engineer

  • Designed and built a RAG chatbot using LangChain and Llama 3.0, incorporating research papers and exercise philosophy to answer user-specific queries; enhanced user interaction and motivation within the Klein Training App for 4000+ users.• Contributed to the strategic development of AI-driven applications to enhance the quality of user UX; developed test vision models for exercise form correction using OpenCV and Mediapipe with potential applications for in-app plyometrics onboarding.• Actively stay up to date on the latest AI research and model developments, applying new insights and techniques to continually improve AI-powered features and solutions. Show less

• 

  The Learning Directed Operating System Expedition

  Oct 2024 - now

  Undergraduate Research Assistant

  • Developing a novel methodology leveraging Retrieval-Augmented Generation (RAG) with Large Language Models (LLMs) to generate reconstructed synthetic code based on sample traces, addressing privacy and scalability limitations in trace datasets, relying on less than 10% of available data.• Implemented a RAG/LLM pipeline inspired by Google's FunSearch paper, evolving a Chroma vector database for future code retrieval and code generation context. Developed a code analyzer, a cloud-based Python environment that runs generated code and uses LLMs to analyze CPU utilization, refining subsequent generations of code and retrieval for enhanced accuracy and performance.• Implemented uniqueness scoring using vector-based code embedding comparisons and Earth mover's distance scoring, reducing redundancy by 40% in generated solutions.• Regularly read and analyze research papers to stay up to date with the latest advancements in AI and model development, integrating new findings into ongoing projects. Show less

• 

  The University of Texas at Austin

  Jan 2025 - now

  Undergraduate Research Assistant

  • Developed an AI-powered tool, MetaCues, to enhance creativity and critical thinking via metacognitive reflection prompts, using LangChain and Ollama open-source LLMs. Improved prompt relevance by 30% through iterative testing and fine-tuning of cognitive-pattern-driven prompt generation.• Designed, built, and deployed a web interface using React, Node, and Flask for the MetaCues application on university Linux servers, creating an intuitive user experience for researchers and participants while ensuring stable access, backend integration, and optimized performance for real-time interactions, scaled to support more than 100 users.• Utilized Docker to store and set up SearXNG for decentralized search and Redis for efficient caching, optimizing application performance and scalability.• Supporting research efforts, including data processing, tool deployment, and drafting findings for submission to CHI, a leading conference in human-computer interaction. Show less