Hugo Ponte

Exploring technology live at youtube.com/@hu-po


I am available for consulting work, email me if you have questions.
Je suis disponible pour des missions de conseil, envoyez-moi un email si vous avez des questions.
Estoy disponible para trabajos de consultorĂ­a, envĂ­ame un email si tiene preguntas.


Experience
2022
Founded a technical consulting company hu-po. I do consulting work in the field of AI&ML, robotics, XR, and more! Livestreamer within the genre of coding and paper reading. Over 500 hours of content mainly in the space of Diffusion models (2D and 3D), LLMs (LoRAs, Quantization, RAG), Humanoid Agents, Gaussian Splats, Speech2Text. Worked on several projects using OpenAI API and Open Source alternatives. Built a community on Discord for ML researchers.
2022
Lead developer on the wandb sweeps product: a popular hyperparameter tuning framework for machine learning research. Production code commits in Python, Go, TypeScript using frameworks like Docker, AWS, GCP, SQL.
2019
Co-founded a company in the Machine Learning software tools market, pioneering Synthetic Data for Computer Vision. Went through Y-Combinator during covid pandemic and raised $1.3 million. Directed product roadmap, managed and hired a technical team of 10 employees. Stack used Python, PyTorch, Unity, C#, Docker, Kubernetes, AWS, GCP.
2018
Designed, trained and analyzed Reinforcement Learning agents deployed on Robots (low-cost servo-based manipulators and quadrupeds) using Tensorflow, JAX, MuJoCo. Team contributor on two publications [1] [2] during the 1-year residency program. Avid attendee of paper reading groups.
2016
First robotics hire built many of the first versions of the robot software stack: camera and vision sensor integrations, Object Detection and Pose Estimation pipelines, robot motion control and planning, Simulation environments for Data Collection and testing. Osaro creates AI software for robotic control and perception.
2014
Owned the system layout (Robotic Path Planning map) of a new prototype robotic warehouse from initial concepts into production surviving Amazon's busy peak season. Improved throughput of the warehouse layout through analysis of Simulation and real-time data using SQL and Python. I also helped organize and judge the Amazon Picking Challenges: international robotic manipulation competitions held in 2015 and 2016.
2011
Graduated with an MS in Robotics, first author publishing work on Control Theory [3] and 3D Structured Light sensors [4]. Robots used C++, Linux, Matlab, Python. Also received undergraduate degrees at CMU in BS in Physics and BS in Mechanical Engineering providing a strong mathematical foundation. Made the dean's list for 4.0 GPA multiple years.

Publications
2020
Hugo Ponte, Norman Ponte, Sammie Crowder, Kory Stiger, Steven Pecht, Elena Ponte
2019
Michael Ahn, Henry Zhu, Kristian Hartikainen, Hugo Ponte, Abhishek Gupta, Sergey Levine, Vikash Kumar
2019
Ofir Nachum, Michael Ahn, Hugo Ponte, Shixiang Gu, Vikash Kumar
2014
Hugo Ponte, Matt Travers, Howie Choset
2014
Hugo Ponte, Max Queenan, Chaohui Gong, Chirstoph Mertz, Matt Travers, Florian Enner, Martial Hebert and Howie Choset