The Allen Institute for AI

The AI Robot Association (AIRoA) is a non-profit organization launching October 1, 2025, with a mission to integrate AI and robotics through open platforms and a global robot data ecosystem. With over 24 member companies including Toyota, Nissan, NEC, Fujitsu, and Mitsubishi Electric, AIRoA operates state-of-the-art infrastructure comprising hundreds of humanoid robots, GPU clusters, high-fidelity simulators, and a global-scale evaluation pipeline. The organization's technical work spans Vision-Language-Action (VLA) models, multimodal learning, tactile-rich manipulation, sim-to-real transfer, and robotics foundation models. AIRoA's core initiative targets collecting one million hours of humanoid robot operation data across hundreds of robots to train what the organization describes as the world's most powerful VLA models. This data collection effort directly supports the development of robotics foundation models and addresses large-scale benchmarking challenges in the field. All datasets, trained models, and benchmarks are made available to the global research community as open resources, reflecting the organization's commitment to advancing robotics technology through collaborative development. The organization provides an open evaluation platform enabling researchers worldwide to test their models on standardized humanoid robots through AIRoA's infrastructure. This platform supports work across multiple technical domains including high-fidelity simulation, sim-to-real transfer validation, and standardized benchmarking. Engineers working with AIRoA's platform will engage with technical stacks including ROS/ROS2, C++, Python, Isaac Sim, MuJoCo, and PyBullet. The organization is led by Tetsuya Ogata and operates with a focus on addressing societal challenges including healthcare efficiency improvements and labor shortages through advanced robotics technology deployment.

We are Shield AI, a US based defense technology company founded in 2015 to protect service members and civilians with intelligent systems. We design and deploy autonomous mission systems for aircraft and unmanned platforms, built around our Hivemind autonomy backbone and a family of mission ready software and hardware modules. Our products include X-BAT and V-BAT aircraft, Hivemind Enterprise, ViDAR and other capabilities built to operate in GPS and comms denied environments. We develop EdgeOS, Pilot, Forge and Commander to enable rapid development, testing and deployment of resilient autonomy across domains. Our approach is aircraft agnostic, battle tested, and focused on human directable autonomy that can scale across platforms and theaters of operation. Our team emphasizes servant leadership and a mission first mindset. Since founding we have grown to over a thousand teammates with offices across the United States and internationally, and we continue to invest in building safe reliable autonomy for the modern battlefield. We collaborate with allies to advance AI pilots powering a safer world, aiming to reduce risk to service members while enhancing decision making and effectiveness in complex environments. Autonomy for the world remains our guiding vision and the greatest victory, we believe, requires no war.

Toyota Research Institute conducts fundamental and applied research across advanced robotics, automated driving, energy and materials discovery, and human-centered AI. The organization operates from a core premise that cutting-edge research should bridge theoretical advances with practical deployment, focusing specifically on technologies that amplify human ability rather than replace it. Their technical work spans both the hardware-software interface in robotic systems and computational approaches to materials science for sustainable energy applications. The research teams tackle several distinct technical challenges simultaneously. In robotics, the focus centers on systems designed to enhance human capabilities in real-world environments. Their automated driving research addresses the complex edge cases inherent in autonomous vehicle deployment, while their energy and materials discovery work applies computational methods to accelerate the identification of sustainable energy solutions. The human-centered AI initiative emphasizes developing AI systems that augment rather than automate human decision-making and physical tasks. The organization maintains a structure that supports cross-disciplinary collaboration between researchers and engineers, creating pathways from breakthrough discoveries to tangible implementations. The work environment prioritizes scientific rigor and empirical validation, with teams expected to navigate the constraints and edge cases that emerge when moving from controlled laboratory settings to field deployment. This operational approach reflects a commitment to translating fundamental research into technologies that address real-world challenges in mobility, sustainability, and human augmentation.

The Robotics and AI Institute tackles fundamental challenges in robotics and artificial intelligence to develop machines capable of perceiving, learning, and interacting with the physical world. Led by Marc Raibert, founder of Boston Dynamics, the organization operates facilities in Cambridge, Massachusetts and Zurich, Switzerland. The institute structures its work around four core research areas: Cognitive Intelligence, Athletic Intelligence, Organic Design, and Ethics - reflecting both technical ambition and deliberate consideration of how intelligent machines should function in society. Research spans the technical stack from perception and machine learning to robot control, locomotion, and human-robot interaction. The organization maintains high-performance computing infrastructure and access to cutting-edge equipment to support development of next-generation robots designed to be more agile and capable than current platforms. Target applications include increasing productivity, removing humans from dangerous work environments, and providing assistive technology for individuals with disabilities. The institute positions itself as a hybrid between academic research labs and corporate R&D organizations - combining the fundamental research orientation of academia with the resources, pace, and focus on deployment characteristic of industry. With substantial backing and infrastructure investment, the organization aims to advance the state of the art in how robots perceive their environment, learn from experience, and physically interact with unstructured spaces. The work directly addresses constraints that currently limit robot autonomy, dexterity, and real-world deployment across industrial automation and assistive technology applications.

Worcester Polytechnic Institute is a private research university founded in 1865, holding R1 research designation as one of only 187 institutions nationwide recognized for very high research activity. With over 70 degree programs spanning undergraduate and graduate levels, WPI operates more than 50 project centers across six continents, enabling hands-on technical work in partnership with communities and organizations worldwide. The institution's defining characteristic is its project-based curriculum, which requires students to apply classroom theory to real-world scientific, technological, and societal problems throughout their academic careers. This approach emphasizes interdisciplinary collaboration among faculty, industry partners, and global communities. The curriculum produces STEM professionals trained in translating technical knowledge into practical deployments rather than theoretical study alone. WPI's research enterprise spans clean energy systems, healthcare innovation, sustainable infrastructure, and technology advancement. The university's operational model centers on use-inspired research - investigations motivated by both fundamental understanding and practical application constraints. For engineers considering positions here, the environment prioritizes hands-on problem-solving, field-tested solutions, and the technical challenges inherent in moving from laboratory concepts to deployed systems across diverse global contexts.

Radical AI operates autonomous laboratories that combine machine learning, robotics, and closed-loop experimentation to discover inorganic materials for aerospace, clean energy, and semiconductor applications. The company's self-driving laboratory system executes synthesis and characterization experiments without human intervention, feeding results back into AI models that screen billions of material compositions to identify candidates for physical validation. This architecture transforms materials R&D from sequential workflows into parallel feedback loops where prediction, synthesis, and performance measurement operate as a continuous integrated system. The technical stack spans Python and C++ for core systems, ROS/ROS2 for robotic control, computer vision and SLAM for laboratory automation, and reinforcement learning for adaptive experimental design. Simulation environments include MuJoCo and IsaacSim for developing manipulation and navigation behaviors before hardware deployment. Infrastructure uses gRPC/protobuf for inter-service communication, OpenTelemetry and Prometheus for instrumentation, with Grafana for monitoring closed-loop experiment execution at scale. With $55M in Series Seed+ funding, the company addresses multi-decade materials challenges in industries where novel inorganic compounds enable next-generation performance - higher-temperature aerospace alloys, more efficient energy conversion materials, and advanced semiconductor substrates. The approach grounds AI prediction in physical constraints: every screened composition must survive automated synthesis protocols, pass characterization under real operating conditions, and demonstrate reproducibility across experiment cycles before qualifying as a discovery.