Journal

A collection of things I've learned, built, and explored.

OCTOBER 19, 2026

Matt Pocock's AI Coding Workflow

I watched Matt Pocock's video on his agent skills workflow. The idea that stuck: make the AI interview you before it writes any code, and never let it review its own work.

OCTOBER 15, 2026

Associate AI Engineer for Data Scientists

I finished DataCamp's Associate AI Engineer for Data Scientists track. Much of it reviewed my master's program, plus the software engineering foundation I didn't have. The associate certification is next.

OCTOBER 14, 2026

Introduction to Testing in Python

I finished DataCamp's Introduction to Testing in Python course, the last one on the Associate AI Engineer for Data Scientists track. It covered pytest, fixtures, four types of tests, and unittest.

OCTOBER 13, 2026

Introduction to Git

I finished DataCamp's Introduction to Git course: creating repos, staging and committing, reading a project's history, and undoing mistakes.

OCTOBER 12, 2026

Software Engineering Principles in Python

I finished DataCamp's Software Engineering Principles in Python course. Without a traditional computer science background, these fundamentals fill a real gap, and MLOps builds on them.

OCTOBER 9, 2026

MLOps Concepts

I finished DataCamp's MLOps Concepts course. Not every ML system needs full automation, and the maturity levels help you pick how much.

OCTOBER 7, 2026

Working with Llama 3

I finished DataCamp's Working with Llama 3 course. Controlling model output through parameters, roles, and prompt structure is what finally clicked.

OCTOBER 6, 2026

Introduction to LLMs in Python

I finished DataCamp's Introduction to LLMs in Python course. LLMs come with real challenges, and evaluation matters because model performance can make or break production.

OCTOBER 6, 2026

Car-ing is Sharing LLM Chatbot Prototype

A DataCamp project I did after Introduction to LLMs in Python: prototyping a car dealership chatbot that runs five tasks with pre-trained Hugging Face models.

OCTOBER 5, 2026

Making a Ghost Mascot with the ip-as-logo Skill

I used a GitHub skill that makes simple, cute character icons to generate a ghost, then put it in the nav and the browser tab where the [T] used to be.

SEPTEMBER 30, 2026

Responsible AI Data Management

I finished DataCamp's Responsible AI Data Management course, covering the six dimensions of responsible AI, data management plans, data source selection, and bias auditing.

SEPTEMBER 29, 2026

Developing Multi-Input Models for OCR

I finished DataCamp's Developing Multi-Input Models for OCR project, building a PyTorch model that classifies scanned insurance IDs as primary or secondary.

SEPTEMBER 28, 2026

Intermediate Deep Learning with PyTorch

I finished DataCamp's Intermediate Deep Learning with PyTorch course, moving past the fundamentals into CNNs, RNNs, and multi-input models.

SEPTEMBER 22, 2026

Jev and the Case for Structured, Confident Outputs

TypeSafe AI's Jev skips free-text generation for structured values with calibrated confidence scores. I want to see how it handles something with clear right and wrong moves: chess.

SEPTEMBER 22, 2026

Explainable AI in Python

I finished DataCamp's Explainable AI in Python course, working through SHAP, LIME, and permutation importance to pry open models that don't explain themselves.

SEPTEMBER 16, 2026

Introduction to Deep Learning with PyTorch

I finished DataCamp's Introduction to Deep Learning with PyTorch course, a refresher on the neural network fundamentals I first learned in grad school.

SEPTEMBER 13, 2026

Practicing Working with Hugging Face

Third course on the way to DataCamp's Associate AI Engineer track. I'd already used Hugging Face in grad school, this filled in the gaps around how I was using it.

SEPTEMBER 12, 2026

Practicing Unsupervised Learning in Python

Second stop on the way to DataCamp's Associate AI Engineer track: clustering, hierarchical clustering, PCA, and NMF, with no labels to check my work against.

SEPTEMBER 3, 2026

Practicing Supervised Learning with scikit-learn

I finished DataCamp's Supervised Learning with scikit-learn course. The exercises were new, but the concepts underneath them came from my Master's program.

SEPTEMBER 1, 2026

What a Real Forecasting Problem Taught Me

I wrote a policy paper on Northern Virginia's data centers built on other people's numbers. I wanted my own, and a real dataset to actually get good at forecasting on instead of a textbook one. Here's what that taught me.

AUGUST 28, 2026

Teaching Myself OCR on Riftbound TCG Cards

OCR, software that reads text out of a photo, was everywhere at my last job, but I'd never built one myself. I picked Riftbound, League of Legends' trading card game, to learn it from scratch, and ran into a second problem I didn't expect: not enough physical cards to test on.

AUGUST 24, 2026

Borrowing Taste: Using GitHub Design Skills to Cut the AI Slop

I don't have a frontend or UX background, so I tried using three open-source Claude design skills to sharpen my portfolio's motion and cards, then ran the same skills on a League of Legends matchup chatbot.

AUGUST 20, 2026

What Building LoL Matchbook Taught Me

Two weeks building a League of Legends app that gives instant matchup advice before your pick timer runs out, and what verifying my own work the hard way taught me.

AUGUST 10, 2026

Completing My M.S. in Data Science at UVA

I've completed my Master's in Data Science at the University of Virginia's School of Data Science, the foundation for what I want to build next.