J. Wilson Peoples Portfolio Website

About Me

I am a Principal AI/ML Engineer at CleanSlate Technology Group, where I am a technical lead on the AI side of the firm's Data & AI practice. I design, train, and deploy machine learning systems that serve millions of predictions per day in production.

I hold a Ph.D. in Mathematics from Penn State (2024), and my research spans two mathematical lives:

Machine learning & applied mathematics: manifold learning, operator estimation, and numerical methods for PDEs on unknown manifolds
Pure mathematics: operator algebras, noncommutative geometry, and K-theory

I am the author of 10 papers and preprints (80+ citations) across these areas, published in venues including the Journal of Machine Learning Research, Foundations of Data Science, and the Journal of Noncommutative Geometry.

Below you'll find highlighted research, engineering projects, and data notebooks. For the complete collection, explore this website using the Portfolio dropdown menu at the top left of each page, or check out my Github.

Highlighted Research

Here are a few highlighted papers from my research. For all of my publications and preprints, along with talk slides, posters, and textbook solutions, view the full Theory page.

Spectral Convergence of Symmetrized Graph Laplacian on Manifolds with Boundary

With J. Harlim. We prove convergence rates for eigenpairs of graph Laplacians toward those of the Laplace–Beltrami operator on manifolds with boundary, the theoretical backbone of diffusion-maps-style manifold learning.

Foundations of Data Science, 2025

Journal arXiv

Radial Basis Approximation of Tensor Fields on Manifolds

With J. Harlim and S. W. Jiang. We develop radial-basis-function methods for estimating differential operators on tensor fields over unknown manifolds identified from point-cloud data, taking us from operator estimation all the way to manifold learning.

Journal of Machine Learning Research, 24(345), 2023

Journal arXiv

Aspects of Noncommutative Geometry of Bunce–Deddens Algebras

With S. Klimek and M. McBride. We define smooth subalgebras of Bunce–Deddens C*-algebras and study their noncommutative geometry: derivations, K-theory, and K-homology.

Journal of Noncommutative Geometry, 17(4), 2023

Journal arXiv

View All Publications

Highlighted Projects

While I'm not able to share details of projects with current clients, below are some personal projects I created while preparing to enter the applied ML/AI space.

Serverless Podcast Transcription Pipeline

In this project, we leverage AWS Lambda, along with asynchronous AWS Transcribe & Comprehend jobs, to create an event based, fast podcast transcription pipeline.

View Project

Real-Time Stock Market Streaming with Spark (Scala)

In this project, we build a cloud native, fully dockerized pipeline streaming real time stock market data across many symbols: orchestrated with Kubernetes, powered by Spark.

View Project

BoundaryDM Manifold Learning Library

This project consists of a scikit-learn style deployment of a novel technique for manifold learning / nonlinear dimensionality reduction, developed by me and my advisor.

View Project

Highlighted Notebooks

These notebooks document the very beginning of my journey into the real-world data space.

Big Data Workflow with Pandas and SQLite3

In this notebook, we demonstrate a workflow using pandas and SQLite3 which scales well with large data.

View Notebook

Implementing Efficient Joins on Mobile App Data

In this notebook, we implement a few different join algorithms, and study their time and space complexity on real world datasets.

View Notebook

Document Search Using Map Reduce

In this notebook, we implement a custom MapReduce framework in python, and use it to create a document search function.

View Notebook