Santiago Restrepo-Garcia

I am a Mechatronics Engineer based in the US. Formerly, I was a Researcher Jr. at Universidad Autonoma de Occidente where I designed and built a heat-press machine for bioplastic materials. Later on, I was a Maintenance Engineer Jr. at the fast-paced tech start-up Kiwibot, where I maintained the mobile delivery fleet at optimal conditions to meet quarters' goals.

Previously I obtained my BEng in Mechatronics Engineering at Universidad Autonoma de Occidente, where I worked on mobile robotics using Micro Air Vehicles with the guidance of Victor Romero-Cano at the Robotics and Autonomous Systems Laboratory. Here, I developed CrazyKhoreia, a low-cost image-based path planner for drone light shows in two modes of operation; and 3DBot, a mobile platform for large-scale additive manufacturing.

I aspire to converge my career into visual perception for mobile robotics and AI, as it has been my main interest over the past few years. Then, I look forward to develop it further in graduate programmes and research laboratories.

Email  /  CV  /  LinkedIn  /  Github

profile photo

Research

I'm interested in computer vision, machine learning, and mobile robotics. I like to build fully connected robotic pipelines where an agent perceives data from its surroundings and takes real-life action from them.

CrazyKhoreia, a robotic perception system for MAV path planning from digital images
Santiago Restrepo-Garcia, Victor Romero-Cano
To be published ...
Project page / Pre-print / Video

An image-based path planner for Micro-Air Vehicles (MAV) based on traditional computer vision algorithms. CrazyKhoreia works as a Python package in two modes of operation where both of them feed from a digital image and output a .csv file. The light-painting mode generates an apparently continuous flight trajectory for long-exposure photography, whereas the multi-drone formation mode coordinates an MAV swarm into calculated positions to resemble the original image.

3DBot, a mobile robot for large-scale additive manufacturing
Oscar D. Diaz-Santos, Diana S. Villaraga, Santiago Restrepo-Garcia

Project page / Video

3DBot is a robotic mobile platform designed from scratch for large-scale additive manufacturing. This robot uses omnidirectional wheels to move freely at any direction, this allows the extruder to have virtually unlimited space across the X and Y axis, while the Z axis is limited by the height of the vertical column. As a fun fact, an early prototype of CrazyKhoreia allowed 3DBot to draw cartoons on a whiteboard using a marker.

Education

  • Mechatronics Engineering BEng. Universidad Autonoma de Occidente. Cali, Colombia.
  • Certifications

  • TensorFlow Developer Certificate [Credential Feb 2024]
  • Courses

  • PyTorch for Deep Learning in 2023: Zero to Mastery by Udemy. A practical course to acquire PyTorch and deep learning skills. [On going]
  • TensorFlow Developer Certificate in 2023: Zero to Mastery by Udemy. An all-hands course practical course oriented on learning deep learning concepts and TensorFlow. [Feb 2024]
  • Learn ROS2 as a ROS1 Developer and Migrate Your ROS Projects by Udemy. A ROS2 migratory course.[Credential Dec 2022]
  • Introduction to Programming in C++ [audit track] by EdX. A C++ beginner course. [Dec 2022]
  • Advanced Programming in C++ [audit track] by EdX. A C++ intermediate course. [Dec 2022]
  • Object Oriented Implementation Using C++ [audit track] by EdX. A C++ intermediate course using OOP. [Dec 2022]
  • Machine Learning by Standford Online, Coursera. An introductory course that covers the basics on Machine Learning. [Credential Dec 2020]

  • Stolen from Jon Barron's website. Updated on July. 13 2024.