Raphael San Andres profile photo

Raphael San Andres

Machine Learning Engineer passionate about developing and deploying innovative ML solutions. Supporting early-stage startups and solving complex ML problems.

Professional Experience

Stealth

Software Engineer (SDE I)

April 2025 - Present
  • Split time between Solutions and Support Engineering and Kubernetes Teams
  • Developed and Managed bidirectional Jira and Kubernetes Operators, easing on-call work by 30%
  • Continuing all responsibilities from previous role as Founding ML Solutions Engineer

Founding ML Solutions Engineer

February 2024 - Present
  • Grew and led a customer support-facing Machine Learning Engineering team from 0 to 5 members, increasing efficiency and reducing support tickets by 20%
  • Managed and supported five 32 to 1600 H100 Kubernetes clusters, ensuring 99.9% uptime
  • Enhanced ML efficiency for 10+ early-stage AI startups by optimizing SLURM and Kubernetes workflows, reducing GPU failure rates by 40%
  • Automated node repair processes, cutting repair times by 90% and saving 50 engineering hours monthly
  • Created and maintained Kubernetes Toolings for debugging and repair, reducing debugging times by 80%

Weights and Biases

Machine Learning Support Engineer

January 2023 - January 2024
  • Debugged and solved 600+ issues from ML Practitioners from OpenAI, NVIDIA, and Microsoft regarding model integrations, LLMs, and local instances
  • Triaged and traced 50+ bugs in the SDK, App, backend, and instances
  • Managed approximately 20 customer requests a day while organizing customer calls, debugging sessions, internal ticket syncs, and personal growth projects (Creating W&B integrations, Bugfix PRs, frontend development, etc)

Education

Masters in Artificial Intelligence (CS)

Penn State — Remote (Grad. May 2025)

Bachelor of Science in Statistics

UCLA — Los Angeles, CA (Grad. June 2022)

Skills

Tools & Platforms

  • Docker
  • Kubernetes
  • AWS SageMaker
  • GCP Vertex
  • Azure
  • Lambda
  • Jupyter
  • DataDog
  • SLURM

Programming Languages

  • Python
  • SQL
  • R
  • C++
  • Go

Libraries & Frameworks

  • TensorFlow
  • Keras
  • PyTorch (Lightning)
  • Ray
  • HuggingFace
  • Jupyter
  • LangChain
  • OpenAI API

Projects

Bird Classifier and GAN

April 2023 - August 2023

  • Created a Bird Classifier with 525 species, achieving an accuracy of 99.81%
  • Developed a DCGAN to create bird images, achieving an average accuracy of 97% on fake images

Q-Learning in Custom OpenAI Gym (Reinforced Learning) - A-I 801

April 2023 - August 2023

  • Designed a custom OpenAI Gym Environment to visualize the agent, rewards, and enemies
  • Implemented 2 separate agent classes with 3, respective action spaces for comparison

Scientific Paper Summarization Tool (Pegasus-X)

September 2023 - Present

  • Collected Scientific paper PDFs (100GB+) from arXiv.org API for training and testing
  • Parsed PDFs into text using PyMuPDF (OCR) and combined with arXiv Metadata to create dataset

Repositories

jaxstats

Active Development

Utilizes JAX to display League of Legends stats locally from your computer. Still under active development.

corium

Work in Progress

Will be a Kubernetes installation of jaxstats.

Capstone (PSU AI 894)

Spring 2024

Repository for PSU AI 894 Capstone Project. Predicts NFL formation based on player positions and coordinates.

Publications

Certifications