🤖 Robotics - Job Openings
Robotics engineers are sought for whole-body control on a humanoid robot at UMA. The position is full-time and located in Paris. Interested candidates should apply through the provided link.
Key Points:
• The job opening is for a robotics engineer to work on whole-body control for a humanoid robot at UMA.
• The position is full-time and located in Paris.
• Candidates should apply through the provided link.
🔗 Resources:
• https://app.dover.com/apply/UMA/5f0e0486-6941-4ccd-b41e-16c611b8f261/ ↗ - Job application link
• https://x.com/LouisLeLay4/status/2094339058050781593 ↗ - Original post URL
🤖 Robotics - Platform Engineer
The platform is building training and evaluation tools for robots. The ideal candidate will have experience with infrastructure (GPU orchestration), robotics, and good people skills.
Key Points:
• The platform is building training and evaluation tools for robots.
• The ideal candidate will have experience with infrastructure (GPU orchestration) and robotics.
• Good people skills are also required.
🔗 Resources:
• https://x.com/Raul_RomeroM/status/2094650544748974112 ↗ - Original post URL
• https://x.com/Raul_RomeroM ↗ - Original poster's Twitter handle
🚀 Manufacturing - New Factory Capacity
Noxmetals has introduced 75,000 sqft of new factory capacity. The metal produced can be found in various applications, including rocket ships, satellites, and nuclear reactors.
Key Points:
• Noxmetals has introduced 75,000 sqft of new factory capacity.
• The metal produced can be found in various applications, including rocket ships and satellites.
🔗 Resources:
• https://x.com/zanehengsperger/status/2094768054152372230 ↗ - Original post URL
• https://x.com/WolfeTyler ↗ - Original poster's Twitter handle
•
Image
🤖 Deep RL - Q-learning to DQN
Q-tables work until the state space becomes too large. DQNs solve this by replacing the table with a neural network that predicts Q-values from high-dimensional states.
Key Points:
• Q-tables work until the state space becomes too large.
• DQNs solve this by replacing the table with a neural network that predicts Q-values from high-dimensional states.
🔗 Resources:
• https://x.com/VizuaraAI/status/2094768484152684944 ↗ - Original post URL
• https://x.com/VizuaraAI ↗ - Original poster's Twitter handle
•
Image
🤖 Computer Vision - 3D Scene Graphs
Generation of High-Level Concepts in 3D Scene Graphs via Autoregressive Diffusion is a research paper that explores the use of autoregressive diffusion for generating high-level concepts in 3D scene graphs.
Key Points:
• The research paper explores the use of autoregressive diffusion for generating high-level concepts in 3D scene graphs.
• The paper is titled "Generation of High-Level Concepts in 3D Scene Graphs via Autoregressive Diffusion".
🔗 Resources:
• https://arxiv.org/abs/2608.28733 ↗ - Research paper URL
• https://x.com/OWW/status/2094768295253573954 ↗ - Original post URL
• https://x.com/OWW ↗ - Original poster's Twitter handle
•
Image
🤖 Robotics - Simulate-first Method
Hugging Face's Microduck is a $399 robot that can be trained in a simulation environment and then run on the hardware. The same policy can be retrained and ported to new engines, driven by language models.
Key Points:
• Hugging Face's Microduck is a $399 robot that can be trained in a simulation environment and then run on the hardware.
• The same policy can be retrained and ported to new engines, driven by language models.
🔗 Resources:
• https://x.com/codecopenflow/status/2094767381478576355 ↗ - Original post URL
• https://x.com/codecopenflow ↗ - Original poster's Twitter handle
🤖 Robotics - Collaborative Research and Development
AIRoA is recruiting collaborative research partners (academia and industry) to advance the initiative for robot data collection using "YUBI."
Key Points:
• AIRoA is recruiting collaborative research partners (academia and industry) to advance the initiative for robot data collection using "YUBI."
• The initiative aims to collect data from robots using "YUBI."
🔗 Resources:
• https://x.com/airoa_jp/status/2094621286496633182 ↗ - Original post URL
• https://x.com/mkftr ↗ - Original poster's Twitter handle
• https://x.com/airoa_jp ↗ - Original poster's Twitter handle
🤖 Robotics - Build JEPA from Scratch Bootcamp
Vizuara's Build JEPA from Scratch bootcamp is a course where participants will learn JEPA, I-JEPA, V-JEPA, Vision Transformers, representation collapse, and action-conditioned prediction by building the system themselves.
Key Points:
• The bootcamp is a course where participants will learn JEPA, I-JEPA, V-JEPA, Vision Transformers, representation collapse, and action-conditioned prediction.
• Participants will build the system themselves.
🔗 Resources:
• https://x.com/VizuaraAI/status/2094766608829022681 ↗ - Original post URL
• https://x.com/VizuaraAI ↗ - Original poster's Twitter handle
•
Image
🤖 Robotics - Open-source Robot
The author is building an open-source robot that can move autonomously through industrial environments. The first real architecture is starting to take shape.
Key Points:
• The author is building an open-source robot that can move autonomously through industrial environments.
• The first real architecture is starting to take shape.
🔗 Resources:
• https://x.com/Lf2p1/status/2094765627944640588 ↗ - Original post URL
• https://x.com/Lf2p1 ↗ - Original poster's Twitter handle
•
Image
•
Image
•
Image
•
Image
🤖 Robotics - Robotic Vision Systems
Robotic vision systems are stretched between the constraints of heat, cost, resolution, latency, and Wi-Fi bandwidth. The author wants a camera with a sort of fovia of high resolution that automatically points itself at things that need more precision.
Key Points:
• Robotic vision systems are stretched between the constraints of heat, cost, resolution, latency, and Wi-Fi bandwidth.
• The author wants a camera with a sort of fovia of high resolution that automatically points itself at things that need more precision.
🔗 Resources:
• https://x.com/VMises76153/status/2094749388782350597 ↗ - Original post URL
• https://x.com/VMises76153 ↗ - Original poster's Twitter handle
