Skip to content

Aligned MachinesTraining AI on human cognition at scale

Human cognition is AI's missing training modality

AI learns from what people write and build, but misses how they perceive or decide. That implicit cognition is what successfully guides us through the complexity of the real world. We turn it into data, rewards, and signals for reasoning machines.

A person wearing the Zypher EEG headset

Capture

Built for large-scale data collection, Zypher EEG hardware captures brain activity from users during real-world tasks with minimal setup times and high mobility. The output is time-synchronized EEG data ready for decoding.

Learn more

Built for AI that moves, collaborates, and decides.

Human cognition provides training and feedback signals for embodied systems, autonomous driving, and personal AI.

  • Operator in tracking glasses teleoperating a dual-arm robotic rig in the Zander Labs labPhysical AI

    Train embodied systems on human intent

    Reward and correction signals derived from operator cognition for teaching real-world manipulation, teleoperation, and shared autonomy.

  • In-cabin cognitive reactions for improved autonomous drivingAutomotive

    In-cabin cognitive reactions for improved autonomous driving

    Continuous ground-truth passenger and driver cognitive states to validate actions of autonomous driving systems.

  • Models that align to the human in the loopPersonal AI

    Models that align to the human in the loop

    High density and multidimensional feedback streams for model and agent alignment to move towards the ultimate personal assistant.

Guided by Curiosity. Building for Human-Aligned AI

We believe in a world where machines support humans and act in accordance with them. We are convinced that using human cognition as a training signal creates a direct path to AI and Human Alignment.

The Zander Labs team

Build with us

Building the next generation of AI and autonomous systems? Let's build together.

Contact us

Latest writing from the lab.

See all publications