Looking for a job

Luis Gomes

@luisgomes1

Software engineering researcher at Carnegie Mellon University building agentic systems aligned with human intent, including visual code tools that cut coding by 49%.

Portugal

Message

At Carnegie Mellon University, I research how agentic systems can act in alignment with human intent. I built a Visual Code Assistant in VSCode that turned ML workflow sketches into Jupyter Notebooks with 79% accuracy for high-level structures, reducing the amount of coding required by 49%.

At Singapore Management University, I built VisDocSketcher, a multi-agent system that generated visual documentation from source code and improved on its baseline by 39.8%. I also designed AutoSketchEval, which scored output quality across 1,000 Jupyter notebooks without ground-truth references.

My research includes peer-reviewed software engineering work at ICSE, ASE, and FORGE, as well as controlled studies with developers from industry. I’ve also worked on ML prediction models and developer tools, and taught courses in ML in Production, software design and testing, and operating systems.

Experience

Work history, roles, and key accomplishments

Visiting Researcher

Singapore Management University

Jan 2025 - Nov 2025 (10 months)

Built VisDocSketcher, a multi-agent system for generating visual documentation from source code, achieving 39.8% improvement over baseline. Designed AutoSketchEval, a reference-free evaluation framework with AUC > 0.87, scaling evaluation across 1,000 Jupyter notebooks.

Multi Agent Systems Static Analysis Evaluation Frameworks Jupyter Notebooks Autoencoder

Data Science Intern

Hubs

Mar 2021 - Jul 2021 (4 months)

Developed visual dashboards in Python to improve internal communication on price predictions and explained ML concepts to team members. Increased prediction accuracy by 3% for ML models through hyperparameter optimization and feature engineering.

Python PyTorch Hyperparameter Optimization Feature Engineering Data Visualization

Junior Software Engineer

Instituto Superior Tecnico

Aug 2020 - Jan 2021 (5 months)

Developed ML architectures for regression and anomaly detection with TensorFlow to predict crowded areas on campus using Wi-Fi data. Collaborated with back-end and front-end teams to design a mobile app for 12,000 students and 1,000 staff, implementing data structures with Kafka, MongoDB, and Python.

TensorFlow Kafka MongoDB Python regression Anomaly Detection Mobile Application Developer

Education

Degrees, certifications, and relevant coursework

Carnegie Mellon University

Ph.D., Software Engineering

2021 - 2026

Ph.D. in Software Engineering as part of the CMU Portugal dual degree program, with relevant coursework in ML in Production, Intro to ML, and Human-AI Interaction.

University of Porto

Ph.D., Computer Science

2021 - 2026

Ph.D. in Computer Science as part of the CMU Portugal dual degree program, with advisors from CMU, Google, and Meta.

University of Minho

BSc + MSc, Computer Science

2015 - 2021

BSc and MSc in Computer Science with a focus on Intelligent Systems and Data Science, including coursework in Neural Networks, Advanced Computing, and Big Data Management.