cv

Basics

Name Angel Ollin Patricio
Label Student Researcher
Email aopatric@mit.edu
Phone +1 (323) 627-0056
Url https://aopatric.github.io/
Summary An East Los Angeles-born researcher pursuing foundation world models and learning algorithms.

Work

  • 2025.01 - Present
    Undergraduate Researcher
    MIT CSAIL
    Working in Prof. Armando Solar-Lezama's group on a benchmark for neurosymbolic reasoning in the domains of planning and perception.
  • 2024.9 - 2025.2
    Undergraduate Researcher
    MIT Media Lab
    Worked on the Camera Culture group's Project SONAR to develop a cross-platform framework for decentralized training of machine learning models through data-secure aggregation (FedAvg, FedProx, etc.).
  • 2024.04 - 2024.09
    Undergraduate Researcher
    MIT IMES
    Spent a summer at IMES working on a project to adapt language models for single-cell RNA sequencing data. Trained on >30M unique cell profiles using up to 6x A100 machines through huggingface accelerate and vast.ai.

Education

  • 2022.09 - 2026.05

    Cambridge, MA

    B.S.
    Massachusetts Institute of Technology
    Artificial Intelligence
    • Machine Learning
    • Real Analysis

Skills

Tools
Python, Julia
Numpy, Pandas
Scikit-Learn, PyTorch
Huggingface, Vast.ai
Langchain, Ollama
Topics
Probabilistic Machine Learning
Reinforcement Learning
Deep Learning
Deep Reinforcement Learning
DevOps
Git, Github
Docker, Kubernetes
AWS, Azure
Linux, Bash

Languages

English
Native speaker
Spanish
Native speaker

Interests

ML
Reinforcement Learning
World Modeling
Learning Algorithms

Projects

  • 2025.06 - 2025.06
    OBRAG
    A cli-based RAG agent connected to your Obsidian vault.
    • Python
    • Langchain
    • Obsidian
    • RAG
    • CLI
  • 2025.03 - Present
    GPD
    A novel decoding approach for dynamic compute allocation at inference time and reduction of compute cost.
    • Python
    • PyTorch
    • Knowledge Distillation
    • Reinforcement Learning
    • Information Theory
  • 2025.06 - Present
    notebook-ml
    A collection of notebooks for machine learning.
    • Python
    • Jupyter
    • Machine Learning
    • Deep Learning
  • 2025.06 - Present
    rolling-notes
    A repo full of my notes and problem solutions in different ML domains and readings.
    • Jupyter
    • Probabilistic Machine Learning
    • Deep Learning
    • Markdown
    • Probability