Bjarte Sunde is a founder and AI/robotics engineer with 13 years of hands-on experience building products from embedded systems to cloud-deployed ML. He has led cross-disciplinary teams at startups like Bumblebee Spaces and Normal Robotics Lab, shipping computer vision, sensing, and HVAC control systems into production and mass-manufactured hardware. An open-source contributor, his popular early-stopping PyTorch package has attracted thousands of users and demonstrates his knack for creating pragmatic tooling that fills real gaps for ML engineers. Early hacker instincts—hacking an iPod at twelve and building a temperamental chess robot—still inform his curiosity for playful, practical prototypes that become robust products. Based in Dublin, he’s now focused on building speaking.app, an AI speech coaching startup that applies his experience in audio, ML, and product design to help people communicate better. He blends hardware fluency, production ML experience, and a startup-founder mindset to move ideas quickly from prototype to market.
13 years of coding experience
4 years of employment as a software developer
Bachelor of Engineering - B.Eng., Automation Engineer Technology/Technician, Bachelor of Engineering - B.Eng., Automation Engineer Technology/Technician at Western Norway University of Applied Sciences
Nanodegree, Deep Learning, Nanodegree, Deep Learning at Udacity
Electro Foundation, and Automation Advanced Courses I and II, Electro Foundation, and Automation Advanced Courses I and II at Sotra vgs avd. Bildøy
Music Performance, General, Music Performance, General at Voss Folk High School
Higher Education Entrance Qualification, Higher Education Entrance Qualification at Akademiet
Contributions:1 release, 28 commits, 47 PRs in 1 year 7 months
Contributions summary:Bjarte implemented an `EarlyStopping` class for PyTorch, enabling training termination based on validation loss. They added essential features like saving model checkpoints and incorporating a verbose argument for monitoring. Furthermore, the user created an example Jupyter notebook demonstrating the use of the `EarlyStopping` class with the MNIST dataset for early stopping in a machine learning model. A small bug fix was also implemented to improve accuracy.
Contributions:35 commits, 20 pushes, 1 branch in 9 days
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.