Library
Everything here is the real document: the textbooks are the official course materials for their Dartmouth classes, and the theses are the filed versions. Click any cover to read.
Textbooks
Written as Head TA; adopted into the Dartmouth curriculum.
Mathematical Foundations of Machine Learning254ppENGS 96, Dartmouth · 2025The mathematics a machine-learning course assumes you already know, written so you actually do: linear algebra, calculus, probability, and optimization built up from first principles.
Game-Theoretic Design, Learning, and Engineering214ppENGS 102, Dartmouth · 2025Games, equilibria, mechanism design, and learning in strategic settings, from matrix games through multi-agent learning.
Principles of Machine Learning280ppENGS 106, Dartmouth · 2026The graduate ML core: estimation, regression, kernels, neural networks, unsupervised methods, and the theory that connects them. The course I took as an undergraduate and later ran as Head TA.
High-Dimensional Sensing and Learning269ppENGS 109, Dartmouth · 2026Compressed sensing, sparse recovery, and low-rank models: how structure makes high-dimensional problems tractable.Theses
Three theses across engineering and music.
MODULO: A Generative Audio Workstation for AI-Native Music Co-Creation360ppM.S. Engineering Sciences · 2026What a DAW becomes when generation is native to it. Five user studies with working musicians; committee-signed final.
Tokenized Curriculum Learning for Multi-Effect Neural Restoration of Full-Mix Music240ppHonors thesis, Engineering Sciences · 2025A 1.08B-parameter token-domain U-Net that jointly reverses five audio degradations in EnCodec's token space. Sigma Xi distinction.
My Life in Sound52ppHonors thesis, Music · 2025The written half of a double honors thesis in music: a self-composed, self-engineered multi-genre album and the story of the ear behind it.Papers
Research writing across generative audio, agents, search, and privacy.
Parallel Harmonic Exploration for Musical Accompaniment with MODULOPDFwith P. Chin, E. Santos, N. Singh · 2026The MODULO study paper: N=25 within-subjects, halved time to a satisfactory accompaniment, tripled harmonic vocabulary, unanimous preference.
Long-Form Speech Reconstruction from Gradients in Federated ASRPDFfirst author, with M. N. Bui and P. Chin · 2025First demonstration of long-form speech reconstruction through CTC loss, exposing speaker-identity leakage in federated learning.
DataGo: Retrieval-Augmented Recursive SearchPDFteam of six · 2025A retrieval layer over KataGo's public networks that reuses expensive analyses across games; wins 90% of stress positions with no retraining.
Unmixing the Music: Structured Recovery for Source SeparationPDFsolo · 2026Robust PCA, NMF, K-SVD, and an unrolled deep sparse coder on one STFT front end, benchmarked against oracle masks on MUSDB18.
Operating-Point Control for Delegated Evidence AcquisitionPDFwith A. Mattar · 2026When should an agent escalate to a human? A prediction-market monitor tuned along its precision-recall operating frontier.
Agentic AI: From Tools to TeammatesPDFfirst author, team of seven · 2025A survey and synthesis of agent architectures: planning, memory, tool use, and what separates a tool from a teammate.
Belief-State Portfolio Allocation as a POMDPPDFsolo · 2026A decision-theoretic trading overlay, reported honestly: the simple baseline wins on Sharpe, and the paper says so.