Case Studies
These case studies show how Yong Qian approaches Physical AI product leadership: defining the right robot use case, building the data and deployment loop, judging field readiness, and turning complex robotics, AI, and industrial technology into products customers can trust.
Physical AI Product Leadership: Humanoid Robots, Data Loops, And Elder-Care Robot Concepts
Yong’s Physical AI work connects humanoid robotics, embedded AI, data infrastructure, and real-world product strategy. The common question is not simply “Can the robot move?” It is: what should the robot do first, what data does it need, where should intelligence run, and what customer workflow makes the concept worth building?
Situation: At Spirit AI, Yong led product vision, strategy, and roadmap work for next-generation humanoid robots. In stealth startup work, he translated customer needs into product direction across IoT, SaaS, AI, and robotics, including humanoid robots and data-collector platforms. With AroOne / 77z, he worked around a privacy-first elder-care companion robot concept combining hardware, embedded intelligence, and on-device AI.
Decision: Connect hardware capability, AI readiness, data-loop design, privacy constraints, care scenarios, and customer value into product direction.
Outcome: Clearer first use cases, task boundaries, autonomy expectations, and data-collection strategy for Physical AI products.
Executive signal: Yong can operate before the Physical AI product is obvious, bringing senior product judgment to the middle between research capability, robot hardware, real users, and commercialization strategy.
Field-Readiness Judgment: Walking Away From A $7M Robotics Deal
Yong’s writing about walking away from a $7M robotics deal explains a core leadership trait: disciplined skepticism. The opportunity was commercially attractive, but the deployment requirements demanded reliable autonomous quadruped performance across unpredictable terrain, weather, endurance, perception, and support conditions.
Situation: The project required robotics capability beyond what the available platforms could responsibly deliver in the field.
Decision: Assess technology readiness, deployment risk, team credibility, and long-term customer trust before accepting the deal.
Outcome: Declined the deal rather than committing to a robotics deployment that could fail outside controlled conditions.
Executive signal: Yong’s robotics judgment is grounded in field deployment, not hype cycles. For Physical AI companies, this is the difference between impressive demos and products that customers can trust.
Read the $7M robotics deal note
Global Automotive Robotics Deployment
Automotive deployment creates pressure that lab demonstrations do not: production schedules, safety standards, quality systems, customer acceptance, penalties, supplier coordination, and long support cycles. Yong’s background across Stäubli, Frimo, and automotive programs gives him a practical lens on what automation systems must survive.
Situation: Global automotive programs require robots, PLCs, vision systems, fixtures, operators, safety teams, quality teams, and suppliers to work together under production pressure.
Decision: Treat robot value as integration value: uptime, safety, serviceability, acceptance, process stability, and customer workflow.
Outcome: Deployment experience across high-pressure automotive environments and global OEM/supplier programs.
Executive signal: Yong understands what separates a working prototype from an accepted production system. That field reality is directly relevant to humanoid robotics and Physical AI deployment.
Read the robotics integration note
Industrial-To-Cloud Platforms: JIEQI And OmniEdge
Before Physical AI became a market phrase, Yong built products around the infrastructure problems robot and machine teams already had: remote access, secure connectivity, diagnostics, distributed devices, and industrial service workflows.
Situation: Factory engineers and machine vendors lost time and money because expert support often required travel. JIEQI / Jaybox turned field-service pain into a hardware + SaaS product for secure access to PLCs, robots, industrial vision, and production equipment. OmniEdge extended this experience into open-source P2P mesh VPN and edge networking SaaS, growing to 7K+ users across 26 countries.
Decision: Build the connectivity layer that lets distributed industrial devices, robots, and edge systems become serviceable products rather than isolated machines.
Outcome: Industrial IoT and mesh networking products with hardware/software depth, global reach, and exits.
Executive signal: Yong can build the infrastructure side of Physical AI: device identity, remote access, edge-to-cloud orchestration, developer experience, and commercial SaaS adoption.
View JIEQI / Jaybox View OmniEdge
Cross-Domain AI Commercialization: DeepFashion
DeepFashion shows Yong’s ability to move beyond his original robotics base and commercialize AI in a different domain. The product started from fast market validation around fashion design workflows and grew into an AI SaaS product using custom generative models, GPU infrastructure, and practical design-tool UX.
Situation: Fashion designers needed AI tools that preserved brand identity, reduced repetitive work, and made generative workflows usable by non-technical creative teams.
Decision: Focus on practical creative workflows: personalized model training, design generation, collection management, and user-facing AI tools.
Outcome: Built, launched, grew, and exited an AIGC SaaS product in a domain outside Yong’s original robotics base.
Executive signal: Yong can cross domains quickly, validate markets, build with modern AI tooling, and commercialize AI products under real constraints.