Omar U. Florez
San Francisco, California, United States
Summary
👤
SeniorOmar Florez is a visiting senior machine learning researcher based in San Francisco with eight years of experience delivering scalable AI systems. He currently leads the distributed pre-training of LatamGPT, a 70B-parameter foundation model built from scratch for Latin America, with a dataset spanning parliamentary records, library collections, and historical archives, and he is releasing training and inference code and datasets to promote reproducibility and regional AI innovation. Previously, as a Senior ML Research Scientist at Twitter Cortex, he pre-trained a multilingual model incorporating social similarity and built a real-time health-score pipeline for tweets across 69 languages, along with a compact distilled mBERT deployed in production. His career spans roles at Intel Labs, IBM, Capital One, and advisory positions shaping Peru's AI National Strategy, reflecting a strong blend of research excellence and cross-industry impact. An entrepreneur at heart, he co-founded PrimerosPuestos.org and has a track record of translating cutting-edge ML research into practical, scalable deployments that empower students, researchers, and startups in Latin America and beyond.
8 years of coding experience
Github Skills (22)
valence10
predict9
python9
machine-learning9
deep-learning9
tensorflow8
tweets8
data-science7
neural-network7
sentiment-analysis7
sequence7
nlp6
lstm6
natural-language-processing5
deep-neural-networks5
Programming languages (3)
Github contributions (5)
omar-florez/scratch_mlp
Oct 2017 - Jan 2020
Contributions:141 pushes, 2 branches, 3 comments in 2 years 3 months
omar-florez/learning_emotions_with_emojis
Aug 2018 - Jan 2020
Predict human emotions in tweets by mapping emojis into the Valence-Arousal space (Russell, 2005). LSTM models the sequence learning of words within the tweets.
Contributions:4 pushes, 1 branch in 1 year 5 months
sequencepredictrussellvalencesentiment-analysis
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.