Ethan Villalovoz

Sacramento, California, United States

Master's student in Computer Science at Georgia Tech. My research interests lie at the intersection of robot learning, world modeling, and human-aligned decision making.

I am always open to connecting. Please feel free to reach out!

Resume CV

News

12/2025

New paper on arXiv on Bayesian prompt optimization for LLM-based code generation.

07/2025

Admitted to the Georgia Tech MSCS program! I’ll be starting in Spring 2026.

06/2025

Gave an alumni talk for the WSU MARC & MIRA program.

Experience

Microsoft

May 2026 - Jul 2026

Software Engineer Intern

Commerce and Ecosystems.

Washington State University

Jan 2024 - May 2025

Undergraduate Research Assistant

Developed and evaluated a Bayesian optimization framework for prompt-based large language model code generation.

Carnegie Mellon University

Jun 2024 - Aug 2024

Robotics Institute Summer Scholar

Developed a hierarchical reward learning framework with Bayesian inference and interactive clarification dialogues.

Google

May 2023 - Aug 2023

Software Engineering Intern (STEP)

Developed scalable C++ and SQL analytics pipelines and interactive dashboards that optimized internal data workflows.

Oregon State University

Jun 2022 - Aug 2022

NSF REU Fellow

Designed and implemented geometric motion primitives and interactive deployment tools enabling expressive multi-robot behaviors.

Publications

View All

An Exploratory Study of Bayesian Prompt Optimization for Test-Driven Code Generation with Large Language Models

S. Tomar, A. Deshwal, E. Villalovoz, M. Fazzini, H. Cai, J.R. Doppa

arXiv, 2025

Explores Bayesian optimization as a principled approach to automated prompt search for large language model–based code generation. Demonstrates sample-efficient improvements in functional correctness over strong prompting baselines on the HumanEval+ benchmark.

Paper

Projects

View All

Self-Driving Car: Behavioral Cloning in the Udacity Simulator

End-to-end CNN (NVIDIA architecture) that predicts steering from front-camera images to autonomously drive the Udacity simulator. Includes balanced/augmented data pipeline, real-time inference via Flask + Socket.IO, and reproducible training.

Computer VisionDeep LearningTensorFlowAutonomous Driving

Code

ClearBill.AI: Explaining Medical Bills with AI and RAG

An AI-powered chatbot that uses Retrieval-Augmented Generation (RAG) with Astra DB, LangChain, and Hugging Face’s Llama-3.1-8B-Instruct to help users understand medical bills with clear, context-aware responses.

RAGLLMsNext.jsVector Databases

Code

Skills

Python

C/C++

SQL

PyTorch

React

Next.js

FastAPI

OpenCV

Hugging Face

Docker

AWS