Darwin Bautista
Staff Data Scientist
Metro Manila, Philippines
Summary
👤
Senior🎓
Top SchoolDarwin Bautista is a Staff Data Scientist with 18 years of software engineering experience and over six years focused on AI across vision and language domains, currently building production-grade AI at Maya. He blends low-level systems knowledge (kernel, microcontrollers) with full-stack development and product-facing ML, having shipped on-device vision models and large-language-model integrations for real-world fintech and KYC systems. A proven technical leader and mentor, he has led cross-functional teams, enforced CI/CD best practices, and upskilled dozens of engineers in distributed projects. His research credentials include a first-author ECCV paper and an open-source PARSeq scene text recognition model that reached #1 on Papers With Code and powered top teams in ICDAR 2023. At Samsung he moved from individual contributor to hybrid technical lead roles, producing patent-grade inventions and delivering global chatbot and telecom orchestration systems. He combines rigorous academic research, practical production deployments, and a knack for reproducing and industrializing cutting-edge ML papers.
18 years of coding experience
12 years of employment as a software developer
Master of Science - MS, Master of Science - MS at University of the Philippines DilimanPhilippine Science High School - Cagayan Valley Campus
English, Filipino
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (5)
computer-vision10
pytorch10
python10
onnx9
machine-learning9
Programming languages (8)
Github contributions (5)
baudm/parseq
Jul 2022 - Jan 2023
Scene Text Recognition with Permuted Autoregressive Sequence Models (ECCV 2022)
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 51 commits, 5 PRs in 6 months
Contributions summary:Darwin primarily contributed to improving the scene text recognition models within the repository. Their work involved fixing configuration issues for ViTSTR, adding features to the `BaseTokenizer` class, updating dependencies, and fixing type hints for Python 3.7. They also made refactoring changes, including adjustments for TorchScript compatibility and the incorporation of label preprocessing options.
pytorchoptical-character-recognitionvision-transformersequencetext-recognition
baudm/vaegan-celebs-keras
May 2018 - Jul 2019
Contributions:47 commits, 37 pushes, 5 branches in 1 year 2 months
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.