01
Capture
Record actions, workflows, and operating states in real kitchen environments.
Physical AI:
A New Recipe for Reshaping the Real World
We began with fried rice, training machines to understand how ingredients change, how heat behaves, and how timing shapes the cooking process.
By learning from human craft, we turn culinary experience and judgment into physical intelligence that can be learned, deployed, and continuously improved.
Cooking combines materials, heat, motion, timing, and human judgment. It is one of humanity’s most refined forms of physical intelligence.
When AI can work with materials, follow complex processes, and create in the real world, the world will no longer be only assembled.
It can be reshaped with skill, memory, care, and culture.
SHEN NONG SHI builds robotic systems that learn from human expertise and perform reliably in complex physical environments.
We start with cooking because it is hard. Real kitchens involve changing ingredients, dynamic heat, precise timing, multi-step workflows, and countless edge cases, making them a powerful proving ground for Physical AI.
From this starting point, we are building the data, models, control systems, and deployment loops that help machines master high-precision tasks in the real world.
From human demonstration to deployable, verifiable, meaningful and scalable robotic systems.
Record actions, workflows, and operating states in real kitchen environments.
Train models to understand physical tasks through demonstrations, teleoperation, and feedback.
Integrate with robotic control systems to complete tasks reliably in dynamic environments.
Evaluate results through taste, aroma, texture, temperature, and consistency.aroma, texture, heat, and consistency.
Turn every deployment and validation cycle into data that continuously drives the next generation of models.
SHEN NONG SHI is built by people who believe that the hardest problems in the real world are worth solving.
This is not a traditional team organized only by hierarchy. It is a mission-driven unit built around real-world tasks.
Here, an engineer might ask the chef, "Why does it have to be flipped at exactly thismoment?" And the chef might ask the model, "How do you know the heat is right?"
This is how we work through intelligence in the physical world.
We work closely with leading research groups in robot learning and human-centered intelligent systems, bringing frontier research into real culinary settings.
#AI #NLP #CV #FinTech #Digital Transformation
# Ph.D. in CS, USC
# B.S. in EE, NTU
# Ph.D. in EE, Purdue University
# B.S. in EE, National Chiao Tung University
# Former Head Chef, Alma Spanish Cuisine
# Former Head Chef, gubami Social
# Family-founded Taiwanese restaurant
“Ching Ching Restaurant”
#Product Management
#Software Development
#Hardware Development
#Artificial Intelligence
#Restaurant Entrepreneur
# Design Trend Analyst
# Scuola Politecnica di Design- Food Design and Innovation
#Home Goods Product Designer
#InteractiveDesign
#DigitalProductManagement
#MixedRealityDevelopment
Each plate of fried rice brings together perception, control, timing, heat, recovery, and consistency.
It is not only measured in a lab, but experienced directly by people.
Featured by NVIDIA founder Jensen Huang as a Physical AI and Robotics use case on the global stage.
Starting from real kitchens, showcasing SHEN NONG SHI's vision for Physical AI to the global innovation and investment community.
One of the earliest startups in NVIDIA’s Physical AI innovation ecosystem.
AI autonomous fried rice core validation, using fried rice as the first complete scenario to validate an AI-driven physical-world model that continuously learns.
If you believe the hardest and most meaningful problems for AI still live in the physical world,
we would love to meet you.
Whether you are a researcher, engineer, chef, designer, operator, investor, or partner, reach out to SHEN NONG SHI.
JOIN US
