John Quevedo • Yale University • Class of 2028
I am a sophomore at Yale majoring in Computer Science and Mathematics. My work spans full-stack products, machine learning, and parallel algorithms, with a consistent focus on clear reasoning, strong technical foundations, and useful systems.
About
I like work that combines engineering execution with real analytical rigor. Computer science gives me the tools to build systems end to end, while mathematics shapes how I think about abstraction, complexity, proofs, and model behavior.
That shows up in the kinds of problems I pursue: understanding how transformers learn to reason, designing parallel algorithms that scale on sparse graphs, and shipping polished product experiences that hold up under real multi-user constraints.
Experience
Yale University • Feb 2026 to Present
Analyzing transformer training dynamics to characterize how attention mechanisms give rise to in-context learning and chain-of-thought reasoning in controlled synthetic settings.
Studying how data distributions, task mixtures, and curriculum schedules influence emergent reasoning behavior during optimization, in collaboration with a CMU PhD mentor.
Yale University • Oct 2024 to Present
Engineered parallel batch-dynamic k-clique counting algorithms for large sparse graphs using low out-degree orientations and work-efficient clique enumeration.
Designed graph primitives including parallel hash tables, adjacency structures, and sparse arboricity-aware representations to improve scalability and memory use.
Management Leadership for Tomorrow • Feb 2026 to Present
Part of a selective 18-month leadership and professional development program focused on coaching, mentorship, and accelerated career preparation for emerging leaders.
Projects
Featured project
Built and deployed a multi-user reading platform where users can search books, manage shelves, write reviews, follow other readers, and track reading activity across the product.
Recommender systems
Built a movie recommendation app with offline matrix factorization training and lightweight new-user inference. Improved top-10 hit rate by 2.2x over a popularity baseline and increased coverage by 9.6%.
ML + time series
Built a five-day realized volatility forecasting pipeline with Gaussian Process regression, predictive intervals, walk-forward backtesting, and a dashboard for uncertainty analysis.
Technical profile
I tend to work on products and models that need both rigor and implementation detail: strong data pipelines, usable interfaces, and defensible evaluation.
Education
I am pursuing a Bachelor of Science in Computer Science and Mathematics at Yale University, expected May 2028, with a 3.83 GPA.
Relevant coursework includes Intermediate Machine Learning, Algorithms, Data Structures, Theory of Statistics, Probability, Discrete Mathematics, Linear Algebra, and Real Analysis.
Skills
Python, C++, Java, JavaScript, TypeScript, SQL.
NumPy, pandas, scikit-learn, PyTorch, TensorFlow, Jupyter.
Git, Linux/Unix, GDB, Docker, PostgreSQL.
Social innovation, entrepreneurship, international travel, and applied technical leadership.
Contact
Reach out for software engineering, quantitative, machine learning, or research conversations.