Ryan Fernandes

Welcome to RyanHub! Easily mistaken for its digital doppelganger, GitHub, I have recently created RyanHub as a place to practice my skills, be creative, and show off my work. There isn't much here at the moment aside from the skeleton, but expect plenty of features and easter eggs to come soon.

I'm a Computer Science and Electrical Engineering student at Yale University with a 4.0 GPA, graduating in May 2028. I'm passionate about AI/ML engineering, open-source software, and building things that make an impact.

ryan@fernandes.us · GitHub · LinkedIn · Resume

Experience

Neural Magic — AI Engineering Intern, vLLM

May 2026 – Present · Boston, MA

• Implemented router-weighted expert pruning compression for large MoE LLMs, reducing VRAM usage up to 50%
• Pruning algorithm reduces VRAM usage up to 50%, saving ~$35,000 per model server on hardware costs
• Building an agent-assisted Kubernetes pipeline to automate multi-day LLM evaluation workflows
• Contributing to vLLM's Speculative Decoding and LLM Compressor repositories (82k stars)

Red Hat, IBM — Software Engineering Intern

May 2025 – August 2025 · Boston, MA

• Created an agentic Docling document editor, yielding a 91% time save on data postprocessing
• Facilitating open-source acquisition by the Docling GitHub organization (35k stars)
• Reduced token costs and latency by 400% through prompt caching and context compression

Education

Yale University

B.S. in Computer Science and Electrical Engineering · Expected May 2028 · New Haven, CT

• GPA: 4.0 (4 semesters, 22 courses), SAT: 1600
• Yale Computer Society Co-President, Club Spikeball Captain
• Coursework: Data Structures & Algorithms, Computer Architecture, Systems & Signal Processing, Digital Systems, Networks, Linear Algebra, Discrete Math, Circuit Analysis & Design

Projects

Yale Computer Society

Co-President, Team Lead of y/labs · Sept 2024 – May 2026

• Ran Yale's largest CS organization — 300 members, 20,000+ users across 7 products
• Drove $19,000 in corporate sponsorships (9.5x increase), managed a $30,000 budget
• Created y/labs, an automated research discovery platform connecting 500+ students and professors

Robotic Die Rotator

April 2026 – Present

• Built an autonomous Arduino system to rotate a playing die to any specified orientation
• Wrote a C++ computer vision algorithm with 98% accuracy on a custom 200-image dataset
• Incorporating into a full assembly line to create pixelated murals from 5,000+ dice

Skills

Python · PyTorch · NumPy · Pandas · Machine Learning · LLM Optimization · Agentic Applications · MCP · Fullstack Web Development · React/TypeScript · Kubernetes · C/C++ · MATLAB · Solidworks · Git