weaklearners

Sean Kim

Looking for a 2027 internship in AI and LLM application engineering.

1 of 3

Who I am

I'm Sean Kim.

I'm studying computer science at WGU. I expect the BS in April 2027 and the MS in October 2027. I build software that runs while I'm asleep. And I keep score on myself.

I'm looking for a 2027 internship building applications on top of AI models. That means AI and LLM application engineering, with backend and full-stack work as the floor.

Work

Three things I built that run. Two have public code.

Overwatch match tracker

I built it to log my own games. Then I used it to kill my own theories.

3,238competitive matches, Nov 2026
Code on GitHub →

nfl-forecast

A prop-ranking experiment that failed twice. The README shows how I found each failure.

726 betsTop third 61.0%, bottom 61.9%
Code on GitHub →

A team of AI agents that works while I sleep

Scheduled agents on three model tiers do research and drafts. Nothing goes out under my name until I say yes.

600+runs, as of November 2026
Private code. Ask me about it.

All code on GitHub → Resume (PDF) →

How I think

Every idea starts out weak. I keep score until it gets strong, or a number kills it.

I keep a list. A belief goes on it the day the data stops agreeing with me.

Gibraltar is my map. 68% win rate. 54.3% over 129 games. The 68% came from a small early sample. Across 30 maps, none of the differences survive the test. 54.3%
I play best at 9 AM and 2 PM. 2 PM is right at my average. 47.6% over 206 games, against 47.5% overall. The story came first. The numbers didn't follow. 47.6%
Noon is my one real bad hour. Still bad. Not proven. 37.0% over 181 games. It passed the test once, in an older count that mixed in practice games. It does not pass now. p 0.0049

Competitive matches only. Numbers as of November 20, 2026.

Things I love

Overwatch

Every match written down since November 2024. The map, the heroes, the result, who killed me, and how the game felt.

3,238competitive matches
24months
47.47%won, sadly

Competitive only. As of November 20, 2026.

Explaining things plainly

If I can't explain an idea to a smart friend who has never met it, I don't understand it yet. Feynman and Sagan set the bar. I use short sentences and a real example before the formula.

Writing

Say hello

sean@weaklearners.com

Or don't write. Just remember the guy who kept score.