years in systems, cloud, and AI infrastructure
publications in networking and optimization
engineers aligned across delivery programs
Jun (Joel) He, PhD
Architecting the standards for the post-deterministic era.
Joel founded OpenKedge around a simple strategic claim: the agentic AI era requires a new systems theory, not just stronger prompts, larger models, or more dashboards.
His work advances Post-Deterministic Distributed Systems (PDDS) as the foundation for reasoning about AI agents, autonomous services, adaptive workflows, and human operators as active participants in distributed systems.
OpenKedge turns that theory into executive-grade control infrastructure: protocols, assurance systems, and governance mechanisms that keep execution authority sovereign, evidence visible, and autonomous action accountable.
A research founder for the control-plane era of AI.
OpenKedge is shaped by the gap Joel has seen across research, global cloud, and large-scale delivery: AI can produce plausible operational intent, but institutions still need deterministic authorization, scoped execution, policy accountability, and durable evidence. That gap is architectural, and PDDS names the theory required to close it.
Standards, theory, and implementation discipline for agentic AI.
PDDS theory and standards
Defining the models needed when distributed systems include adaptive, autonomous, probabilistic, and human-mediated participants.
Sovereign execution control
Separating global reasoning from local authority so autonomous systems act only through policy, identity, evidence, and accountable control planes.
Executive-scale adoption
Moving new theory from research foundation to institutional deployment with the standards, governance, and operating models leaders can trust.
"The next generation of AI infrastructure will not be judged by how much autonomy it grants. It will be judged by whether autonomy can be admitted, bounded, explained, and governed at institutional scale."
This is the PDDS premise: as participants become adaptive and non-deterministic, governance must move from policy documents into the execution architecture itself.
AI output is a proposal until admitted by policy.
Execution authority must remain governed even when models become global.
High-impact autonomy needs standards, not informal trust.
Every important action should leave evidence that executives, operators, and auditors can replay.
Useful where autonomy meets sovereign authority.
The founder perspective is strongest in environments where AI systems may reason globally, but execution decisions must remain locally governed, institutionally accountable, and verifiable by evidence.
Sovereign AI infrastructure
Architectures for institutions that need autonomous capability without surrendering execution authority.
National-scale digital systems
Control-plane patterns for multi-organization environments where authority, accountability, and evidence must cross institutional boundaries.
Executive AI strategy
Advisory work for leaders deciding where AI agents may reason, where they may act, and which standards must govern the space between.
Evidence-backed assurance
Governance architectures where policies, identities, decisions, and execution artifacts become a durable record of institutional control.