Arthur Jakobsson

Robotics Researcher @ CMU

Birds, photography, and computer science.

Hi! I'm a Masters in Robotics student at Carnegie Mellon University working with Professor Jeff Ichnowski. I intend to pursue a PhD in Robotics following my masters and am specifically excited about the topics of dextrous and bimanual manipulation, high-speed manipulation, and creative sensing, all through the lens of ML integration. I hope both to expand on the depth and breadth of my research during my masters.

This summer, I am excited to be a course instructor for Principles of Imperative Computation, a class I have been a TA and head TA for during the last three years. I am passionate about teaching. To learn more about my teaching aspirations and the work I have done please see my page on teaching.

I also love birdwatching, photography, biking, and badminton.

Skills & Expertise

Programming Experience

C, PyTorch, Python, R (+ggplot), C++, Java, Javascript, NodeJS

ML Development Experience

GNNs, CNNs, Diffusion, GANs, RL, image segmentation, one/few-shot learning

Languages

English, Swedish. Elementary: Thai, Japanese, Spanish

Interests & Awards

Photography, Biking, Badminton. Dean's List High Honors (x4), Dean's List (x2). 1st Place Coolest Graphs (CMU Statistics Department) - View Project

Experience

Course Instructor

Principles of Imperative Computation (15-122)

May 2025 - Present

Instructor and lecturer for course after 3 years of being TA & Head TA. As a TA, I pioneered and led development of extra instruction bootcamps for over 1000 cumulative attendants. Developing course infrastructure, managing students & course staff, leading recitations (~40 students).

ML & Visual Manipulation Research

Momentum Lab, CMU

Aug 2024 - Present

Proposed and pursued Honors Computer Science Senior Thesis, advised by Professor Ichnowski. Studying dexterous manipulation of rope and textile with visual input for complex tasks such as table flattening and folding. Joined a team investigating how to leverage sound and visual information to improve contact and hand pose estimation for hand-object interaction.

ML Research

Search-based Pathplanning Lab, CMU

Mar 2023 - May 2025

Engineered postprocessing method using existing heuristic search algorithms to significantly boost ML multi-agent pathfinding success rates. Published & presented oral paper at ICAPS 2024. Using machine learning to generate better and faster results for multi-agent pathfinding (e.g. applicable for pathfinding for robots in warehouses or self-driving cars). Paper accepted to ICRA 2025.

Computer Vision Research

Biorobotics Lab, CMU

Oct 2023 - Nov 2024

Developing methods in computer vision and machine learning to analyze and detect anomalies in few-shot scenarios. Focus on GANs, and diffusion for image segmentation and anomaly detection. Working for ARPA-E Mapping project for robot gas pipe mapping and repair.