cv

Basics

Name Martín Bravo
Label Bachelor of Engineering, Computer Science (Candidate)
Email martin.bravo@ug.uchile.cl
Url https://martinebravo.github.io
Summary Aspiring graduate student in machine learning and AI, focused on robust and privacy-preserving learning systems, with experience spanning research and industry.
Languages English, Spanish, Swedish, Portuguese

Work

  • 2025.11 - Present

    Abu Dhabi, UAE

    Visiting Student
    Mohamed bin Zayed University of AI
    Selected for the first cohort of the Aspire PhD Fellowship program for short-term research with MBZUAI faculty.
    • Research focus on deep neural network sparsity under the supervision of Prof. Samuel Horváth.
    • Working on efficient training and inference techniques for modern deep networks.
  • 2025.06 - 2025.07

    Abu Dhabi, UAE

    Research Intern
    Mohamed bin Zayed University of AI
    Selected among 60 interns out of 2000+ applicants for the fully funded UGRIP research program.
    • Worked on curriculum-based multilingual knowledge distillation for low-resource NLP.
    • Presented final results at the program symposium and received Best Presentation.
  • 2025.01 - 2025.05

    Stockholm, Sweden

    Machine Learning Engineer Intern
    Hopsworks AI
    Built AI-powered tools to optimize and automate ML workflows on Hopsworks.
    • Developed AI-assisted tooling for ML pipeline optimization and large-scale processing.
    • Implemented predictive analytics components to improve decision-making in ML workflows.
    • Designed feature-store-oriented patterns for real-time enterprise ML use cases.
  • 2023.12 - 2024.08

    Santiago, Chile

    Machine Learning Engineer Intern
    SoyMomo
    Developed AI products for child and family safety.
    • Designed and deployed HeyMomo, a child-friendly AI assistant using LangChain, AWS, and Node.js/Express.js.
    • Built and deployed a baby-cry classification model on Azure ML with 92% accuracy and ~2s response time.
    • Created an audio preprocessing pipeline to standardize clips and improve downstream model performance.
  • 2023.09 - 2023.12

    Santiago, Chile

    Student Intern
    National Center for Artificial Intelligence
    Contributed to generative AI applications for cultural experiences.
    • Fine-tuned Stable Diffusion for a museum exhibit with Automatic1111 integration.
    • Built a Flask-based tablet UI to generate personalized AI images for visitors.

Education

  • 2024.08 - 2025.06

    Stockholm, Sweden

    Exchange Student
    KTH Royal Institute of Technology
    Machine Learning
    • Completed graduate-level coursework equivalent to year 1 of the MSc in Machine Learning
    • Joined the TCS research group of Prof. Aristides Gionis and Prof. Sebastian Dalleiger
  • 2021.03 - 2026.06

    Santiago, Chile

    Bachelor of Engineering
    University of Chile
    Computer Science
    • 5.5-year integrated program: foundational coursework (2021-2024), advanced coursework and thesis (2025-2026)
    • GPA: 6.1/7.0
    • Thesis: Lossless Compression for Large Language Models

Publications

  • 2026.01.01
    Robust Federated Clustering under Heterogeneity and Adversaries
    AISTATS 2026
    Clustering distributed and private data is an increasingly important task across domains that handle sensitive information, such as life sciences and clinical research. In federated settings, clustering faces three challenges: heterogeneous client data distributions, adversarial behavior, and strict privacy requirements. Existing approaches often exhibit significant performance degradation under these conditions and fail to return accurate solutions. To overcome these limitations, we introduce a novel federated clustering algorithm that combines client-level differential privacy with Byzantine-robust aggregation at the server, based on a novel efficient and robust clustering procedure. Our method comes with theoretical robustness guarantees, and through extensive experiments on synthetic and real-world data, we demonstrate that it produces high-quality clusters in just a few communication rounds, even in scenarios where state-of-the-art methods fail.

Projects

  • 2025.01 - 2025.12
    OpenAI Agent SDK - Visualization Module
    Designed and implemented the official workflow visualization module for OpenAI's Agent SDK.
    • Implemented dynamic rendering for agent workflow debugging and analysis.
    • First external contributor to the SDK and second-largest contributor at merge time.
    • Feature integrated into official documentation and presented at AI Engineering Stockholm 2025.
  • 2024.01 - 2025.12
    Serverless ML System for IMDb Movie Rating Prediction
    Automated end-to-end ML system for movie rating prediction with data, training, and inference pipelines.
    • Built automated data ingestion, feature engineering, training, and inference pipelines with Hopsworks and GitHub Actions.
    • Developed a web UI to compare predicted IMDb ratings against live IMDb pages.
    • Presented at PyData Stockholm 2025.

Volunteer

Awards

Skills

Programming
C/C++
Python
PyTorch
scikit-learn
JavaScript
Next.js
Tools
Git
Docker
Kubernetes
Hopsworks
Azure
Research Areas
Federated Learning
AI Privacy
Robust Machine Learning

Languages

Spanish
Native
English
Fluent
Swedish
Intermediate
Portuguese
Intermediate

Interests

Research Interests
Federated Learning
Differential Privacy
Byzantine Robustness
Model Compression

References

Prof. Samuel Horváth
Associate Professor, MBZUAI - samuel.horvath@mbzuai.ac.ae
Prof. Gonzalo Navarro
Full Professor, DCC, University of Chile - gnavarro@dcc.uchile.cl
Prof. Sebastian Dalleiger
Assistant Professor, KTH - sdall@kth.se