I work on applied machine learning for scientific problems: models that fuse messy, multi-modal real-world data into predictions scientists can use. At the NMSU KDD Lab, my senior-thesis research uses multi-branch LSTMs to predict net primary productivity across 71,000+ pixels of rangeland from satellite, precipitation, cattle-GPS, and land-cover data. At NASA's Planetary Data System (Atmospheres Node), I develop ELSA, the web application planetary scientists use to generate and archive metadata for large datasets.
Selected work
Research
Predicting Net Primary Productivity with Multi-Branch LSTMs
KDD Lab, New Mexico State University · Sep 2025 – Present
Read the write-up
A multi-branch LSTM that fuses satellite, precipitation, cattle-GPS, and land-cover data to predict net primary productivity across 71,000+ pixels, reaching R² 0.98.
- LSTM
- multi-modal fusion
- temporal encoding
- remote sensing
- time-series regression
Engineering
ELSA: Metadata Tooling for NASA's Planetary Data System
Research Assistant / Software Developer · May 2024 – Present
Read the write-up
A web application that lets planetary scientists generate and archive metadata for large datasets at NASA's PDS Atmospheres Node, including a new service for archiving atmospheric model data.
- ELSA
- Bootstrap 5
- 2FA
- reCAPTCHA
Projects
pantry.ai
1st Place, Desert Dev Lab HackathonBuilt at a hackathon
PWA using Gemini Vision to identify pantry items, estimate expiry dates, and suggest recipes before items expire.
- Next.js
- TypeScript
- SQLite
- Google Gemini API
Airbnb Price Prediction & Feature Analysis
Class project
KNN regression for Airbnb price prediction, with Decision-Tree feature selection identifying the ten most predictive features.
- Python
- Pandas
- Scikit-learn
Compiler Construction
Class project
C-like compiler built with Flex/Bison, AST parsing and a RISC-V code generator handling stack management, recursion, and control flow.
- C
- Flex
- Bison
- RISC-V Assembly