Andrew Dai
Co-founder & CEO at Elorian AI
Palo Alto, California, United States
Summary
🤩
Rockstar🎓
Top SchoolAndrew Dai is a machine learning researcher-turned-founder with over a decade of experience building large-scale language and multimodal models, most recently founding Elorian AI to advance next-generation multimodal intelligence. Previously he spent 14 years at Google/DeepMind leading high-impact efforts including PaLM 2 pretraining, GLaM MoE LLMs, Gemini and synthetic data co-leads, and applied research across Google Health and Search. He combines deep academic training (PhD in Machine Learning from Edinburgh, BA in Computer Science from Cambridge) with hands-on engineering—his open-source contributions to tensorflow/models include substantive fixes and adversarial-text improvements that optimized memory and accuracy. Based in Palo Alto, he blends product-focused leadership as a director-level research scientist with low-level model engineering instincts. Colleagues describe him as someone who moves quickly between strategy and code, able to ship research into production. He brings an unusual mix of enterprise-scale model production experience and entrepreneurial drive to commercialize advanced AI.
10 years of coding experience
15 years of employment as a software developer
Ph.D., Machine Learning, Ph.D., Machine Learning at The University of EdinburghBA, Computer Science, BA, Computer Science at University of Cambridge
Chinese
Stackoverflow
128reputation
7kreached
2answers
2questions
Github Skills (18)
python10
machine-learning10
rnn-model10
n10
tensorflow10
nlp10
model-optimization9
adversarial-learning9
optimisation9
lstm9
optimization9
computer-engineering8
git7
design-patterns6
bazel6
Programming languages (2)
Github contributions (5)
tensorflow/models
Sep 2017 - May 2019
Models and examples built with TensorFlow
Role in this project:
ML Engineer
ML EngineerContributions:33 commits, 27 PRs, 13 pushes in 1 year 8 months
Contributions summary:Andrew primarily contributed to the `research/adversarial_text` module within the TensorFlow models repository, fixing bugs and improving the adversarial text models. The commits involved adjusting model shapes, switching from static_rnn to dynamic_rnn for memory optimization, and correcting accuracy calculations for binary classification. Furthermore, the user added an option for single-label classification and addressed adversarial training issues.
deep-learningtensorflow
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.