NFL decision service for the 2026 season. Calibrated win likelihood and fantasy projections served as static artifacts. The claim is calibration, not beating the market.
in development / Python / DuckDB / LightGBM / Astro / Cloudflare Workers
andres@cmd-lopez:~$ whoami
Andres Lopez
Mechatronics Engineering student at the University of Maryland. I build machine learning systems, reinforcement learning systems, data analytics systems, knowledge graphs, and embedded hardware.
Projects
Projects connected by the tools and systems they share.
Incremental city-builder written from scratch in C++ and Vulkan 1.3. Procedural city generation, road networks, and a save system with offline progress.
in development / C++ / Vulkan / CMake
A life platform for seniors: daily dashboard, medication organizer, family coordination, and a scam-protection shield. Built with a security-first architecture.
prototype / React / FastAPI / SQLAlchemy / Supabase / Twilio
Differential equations toolkit in calculator-constrained Python. Exact rational arithmetic, 921 tests, byte-budgeted builds sized for TI-84 Evo memory.
writeup only / Python
GitHub contributions
A year of commits, pull requests, and reviews on GitHub.
1,464 contributions in the last year · last synced 2026-09-01
1,464 contributions in the last 26 weeks · last synced 2026-09-01
| Month | Contributions | Active days |
|---|---|---|
| Sep 2025 | 0 contributions | 0 |
| Oct 2025 | 0 contributions | 0 |
| Nov 2025 | 0 contributions | 0 |
| Dec 2025 | 0 contributions | 0 |
| Jan 2026 | 0 contributions | 0 |
| Feb 2026 | 0 contributions | 0 |
| Mar 2026 | 3 contributions | 1 |
| Apr 2026 | 0 contributions | 0 |
| May 2026 | 0 contributions | 0 |
| Jun 2026 | 226 contributions | 16 |
| Jul 2026 | 296 contributions | 22 |
| Aug 2026 | 939 contributions | 28 |
| Sep 2026 | 0 contributions | 0 |
Now
- Refactoring my sports prediction work into Sporacle, launching for the 2026 NFL season
- Part time at Skyward IT Solutions, working on government solutions
- Taking Mechatronics Engineering classes at the University of Maryland
- Writing up the data-leakage story from my first prediction models (92.5% to 68.5%, and why that was progress)
Contact
Email is the fastest way to reach me. I read everything.