Jesse Windle

DYDX Capital→

Built a data-driven deal sourcing platform that filters thousands of potential investments down to tens for human review each month.

• Built a system for caching, retrieving, and updating arbitrary web-based data (Typescript, Next.js, Prisma).
• Built a system that does feature extraction, model fitting, and model prediction to build "quality scores" for early-stage startups based on founder demographics. (Features were extracted from either a Postgres database or via NLP and then fed to a scikit-learn or pytorch-based pipeline for modeling.)
• Built a platform to track early-stage startups and their founders. The platform ingests companies and combines rules-based filters, agent-based reasoning, and aforementioned "quality scores" to prioritize startups for human review. (Typescript, Next.js, Prisma, Langchain, OpenAI API.)

Self-employed

Early technical validation for DYDX Capital

• Conducted extensive literature review.
• Completed comprehensive model review to establish baselines, best models, and expected performance.
• Presented findings at DYDX investor day.

Hi Fidelity Genetics / Technologies

Built a completely novel root measurement system that captured full-season, in field, high-throughput, time-lapsed growth.

• Was employee No. 1.
• Conceived and co-invented a device for in-field root growth measurement.
• Built the whole team out, from the initial core to a crew of 13 people. Because we were solving a complex problem, the team had a broad background spanning data science, hardware engineering, software engineering, biology, plant breeding, and agronomy.
• Raised $2M+ in non-dilutive funding (ARPA-E, SBIR).
• Helped develop data systems to manage large-scale plant phenotyping experiments. (Postgres backend with a Flask API.)
• Built the company's entire data science and analytical infrastructure from scratch. (Python pipeline for data ETL, feature extraction, data exploration, and visualization; R for statistical analysis, visualization, hypothesis testing and reproducible results.)
• Developed a novel 3D root growth modeling framework. (Bayesian model recapitulating crown root growth for monocots; fit using Stan.)
• Delivered results to customers. (Customers would come to us to ask questions about root growth. We would manage the experiment and provide the scientific and statistical answers to their questions.)

Duke University

Taught introductory statistics and conducted research in Bayesian statistics.

Duke University

University of Texas at Austin

University of Nebraska - Lincoln

Statistics and machine learning (Expert)

regression and classification, clustering and dimensionality reduction, hypothesis testing, variable selection, state-space models, Bayesian inference, neural networks, tree-based methods like random forests and gradient boosting, Gaussian processes

Statistics and machine learning tools (Expert)

Python, Numpy, Pandas, Scikit-learn, Statsmodels, PyTorch, Lightning, JAX/Flax, R, Stan, Pyro

Systems tools (Expert)

SQL, MongoDB, FastAPI, SQLAlchemy, Next.js, Prisma, Git, Docker / Podman, AWS, Linux

Capturing in-field root system dynamics with RootTracker

Plant Physiology • November 2021

J.J. Aguilar, M. Moore, L. Johnson, R.F. Greenhut, E. Rogers, D. Walker, F. O'Neil, J.L. Edwards, J. Thystrup, S. Farrow, J. Windle, P.N. Benfey. Plant Physiology, 187(3):1117–1130

A tractable state-space model for symmetric positive-definite matrices

Bayesian Analysis • December 2014

J. Windle and C. Carvalho. Bayesian Analysis, 9(4):759-792

The Bayesian Bridge

Journal of the Royal Statistical Society Series B • September 2014

N. Polson, J.G. Scott, and J. Windle. Journal of the Royal Statistical Society Series B, 76(4):713–733

Bayesian Inference for Logistic Models Using Polya–Gamma Latent Variables

Journal of the American Statistical Association • December 2013

N. Polson, J.G. Scott, and J. Windle. Journal of the American Statistical Association, 108(504):1339-1349

rootmodel

The code (R and Stan) supporting the paper Inferring monocotyledon crown root trajectories from limited data

R • Stan • Root modeling

gmmfun

A Python package for fitting distributions via their moment generating function using the generalized method of moments

Python • Statistics

ctgauss

A Python package for sampling from a Gaussian random variable conditioned on a piecewise linear function

Python • Statistics

BayesLogit

An R package for sampling from the family of Polya-Gamma distributions

R • Bayesian statistics

Inferring monocotyledon crown root trajectories from limited data

Manuscript on root modeling methodology

Root modeling • Statistics

Sampling from a Gaussian distribution conditioned on the level set of a piecewise affine, continuous function

Manuscript on HMC sampling within a constrained space

Statistics • Sampling

Efficient Data Augmentation in Dynamic Models for Binary and Count Data

Manuscript on Bayesian analysis of time series data with binary and count observations

Statistics • Bayesian methods

Forecasting High-Dimensional, Time-Varying Variance-Covariance Matrices

Ph.D. Thesis, University of Texas at Austin, 2013

Statistics • Time series