About

I'm a robotics and AI systems engineer. I've built software for humanoid robots in Europe and Taiwan, integrated multimodal AI systems, and I'm now focused on Aware — an effort to give humans and agents a shared, structured environment to collaborate in.

The arc

Humanoids taught me that intelligence needs a body and a world. Working on real humanoid platforms made it obvious that perception, motion, and reasoning only become useful when they're grounded in something shared — a world the robot can refer to, act in, and be corrected against.

AI systems taught me that models are ahead of the environments we run them in. Modern multimodal models are extraordinary, but we still drop them into ad-hoc scaffolding and hope coordination emerges. It doesn't. The missing layer is not a bigger model; it's the structured environment underneath.

Aware is the synthesis. Structured digital environments where humans and AI operate as actors over shared state — a world they can both perceive, act in, and be accountable to. The through-line of all my work is one bet — see the thesis.

Experience

• Aware

  Founder — structured digital environments where humans and AI operate as actors over shared state.

• NEURA Robotics

  AI team — multimodal integration and validation infrastructure.

• Aeolus Robotics

  Humanoid robotics software and AI-enabled human-robot interaction (Taiwan).

What I work on

Robotics

• Humanoid robotics software and behavior systems
• Navigation and manipulation integration
• Simulation, testing, and AI-enabled HRI

AI systems

• Multimodal integration and agent systems
• Runtime orchestration and structured execution environments
• Ontology- and graph-oriented software design

Current direction — Aware

• Structured digital environments and shared world state
• Graph-native runtime abstractions
• Deterministic execution and human + AI coordination

Open source

• aware-sdk

  The current public Aware release. CLI-first distribution covering release and bundle tooling, manifest-driven testing, filesystem indexing, environment contracts, and terminal runtime infrastructure. The OSS rail the broader Graph-OS direction will build on.

• general_manipulation

  End-to-end perception and manipulation exploration using transformer-based approaches for more general robotic manipulation workflows.

• grab_anything

  Computer vision + grasp planning. Awarded 3rd place at the World MoveIt Day Hackathon.

Earlier Aware iterations and agent frameworks: aware · newrail_bot.

Stack

Python, C++, PyTorch, ROS, Docker.

Contact

luis@aware.run · LinkedIn · GitHub

Last updated April 2026.