Data Scientist | Machine Learning Engineer El Paso, Texas
Ph.D. in Data Science. I build compact machine learning models that stay cheap enough to deploy and honest enough to trust — anomaly detection, efficient LLMs, and the monitoring that catches models when they decay.
- LogTinyLLM — compact transformer, 98.83% anomaly-detection accuracy at ~70% lower compute cost than large-LLM baselines (arXiv:2507.11071)
- Drift Arena — live model monitoring: PSI drift detection, McNemar-gated promotion
- The Label Budget — three ML algorithms from first principles, no libraries
Links: Portfolio · Google Scholar · LinkedIn
Python PyTorch SQL R Machine Learning Deep Learning LLMs MLOps