Auralink: A Physical AI Research Reference Architecture
Auralink is a Hyperion-authored Physical AI research agentic software system for EV-charging infrastructure — a safety-relevant, standards-governed domain. It has been evaluated in simulation and bounded environments; it is documented only within its published research boundary. This page uses it as architecture and safety evidence, not as certification, a client engagement or proof of production deployment.
Last reviewed: June 2026
Contents
Auralink is a Hyperion-authored Physical AI research agentic software system for EV-charging infrastructure, built by Hyperion's founder. Its charger OS is aligned to ISO 15118-20 and OCPP. Its implemented architecture provides an inspectable reference for edge-first operation, explicit service boundaries and bounded agent authority. Evaluation is limited to simulation and bounded environments; Auralink is documented within a controlled research boundary. This is engineering evidence, not certification.
The Reference Thesis: A Full-System research architecture in a Safety-Relevant Domain
Most discussion of agentic AI lives in isolated demos and benchmarks. Auralink is different in a narrower, verifiable way: it is a Hyperion-authored research reference for EV-charging infrastructure, evaluated in simulation and bounded environments and documented within a controlled research boundary. That makes it useful for inspecting architecture decisions while keeping its evidence boundary explicit.
The thesis of this page is narrow and deliberate. Auralink is presented as a reference architecture — a worked example of how to architect Physical AI for a standards-governed physical domain — and as safety and architecture evidence. It is not presented as a speed story, a velocity metric, a customer case study or evidence of private repository scale.
Two properties define the reference's design intent. First, sovereignty: the source architecture targets operator-controlled infrastructure and does not require an external cloud in the control path. Second, edge-first continuity: it is designed to preserve bounded local operation when connectivity is degraded or absent. In a physical domain, both are safety-relevant requirements that still require hardware and field validation.
One honesty note, stated up front and again below: Auralink is a Hyperion-authored research reference. It is used here to inspect architecture and safety decisions. It is not an external client engagement, and no external client names or results are claimed anywhere on this page.
What Makes Auralink a Useful Reference
Physical AI research reference system
An implemented architecture for physical charging infrastructure, evaluated in simulation and bounded environments and documented within a controlled research boundary.
Standards-governed domain
EV charging is safety-relevant and governed by ISO 15118-20 and OCPP.
Sovereign infrastructure
Designed for operator-controlled infrastructure; the control-path design does not require an external cloud.
Edge-first / offline-capable
Designed to keep operating under degraded or absent connectivity.
Architecture: Edge-First, First-Party Services, Agents
Auralink's source architecture defines a fan-out of small units intended for independent deployment at the edge, with an agentic layer bounded by explicit authority. The descriptions below state design intent and simulation scope without publishing private inventory or implying physical deployment.
On the agentic layer: the agents assist operation — they sense, reason, and arbitrate actions within bounded authority. The engineering team designs, reviews, and owns the system. The agents do not unilaterally build or ship it; they act inside an envelope the architecture defines.
Implemented research architecture
Implementation statusThe reference implements edge-first decomposition, explicit interfaces and an enforced separation between agent reasoning and deterministic controls. Private repository inventory is deliberately excluded from this public evidence.
First-Party Services
Architecture patternThe system is decomposed into first-party services. This is a fan-out architecture: small, independently deployable units with narrow responsibilities, communicating over well-defined contracts. The service boundary is what makes the system observable, individually testable, and recoverable at the unit level.
Autonomous AI Agents
Agentic layerThe research implementation defines AI-agent roles that assist operation — sensing, reasoning over telemetry, and proposing and arbitrating actions within bounded authority. They do not unilaterally build or ship the system; the engineering team designs, reviews, and owns it. The agents remain inside an envelope that the architecture, not the agent, defines.
Edge-First, Offline-Capable Operation
Edge-firstAuralink's source architecture targets placement at or near charging infrastructure and continuity under degraded or absent connectivity. Edge-first is a safety and continuity requirement for later validation, not evidence of deployment. The design does not require a distant-cloud round-trip for local decisions.
Mapping to the Physical AI system architecture: Sense → Orchestrate
Hyperion frames Physical AI as six verbs: Sense · Connect · Compute · Reason · Act · Orchestrate. The mapping below shows how Auralink's source architecture expresses the intended responsibilities of each layer. It describes design intent and simulation scope, not observed operation on physical charging infrastructure.
Sense
The source architecture defines edge acquisition of telemetry from charging hardware, power electronics, session state, and environmental signals. Physical signal acquisition remains a hardware-integration and validation gate.
Connect
The source architecture defines open-protocol interfaces and aligns the charger OS to OCPP and ISO 15118-20. Current device and cross-vendor interoperability remain separate validation gates.
Compute
The source architecture places computation on operator-controlled edge infrastructure and decomposes workloads into first-party services close to the hardware they would serve. The design does not require an external cloud control plane for local logic.
Reason
The source architecture constrains an agentic layer to reason over sensed or simulated state, diagnose conditions, and propose or arbitrate actions within bounded authority. Behaviour on physical systems remains subject to qualified integration and validation.
Act
The source architecture defines controlled interfaces through which proposed actions could reach charging infrastructure. Physical actuation, protocol interoperability, and safety behaviour remain separate qualification gates.
Orchestrate
The source architecture defines an orchestration layer intended to coordinate lifecycle, deployment, versioning, and recovery across edge sites, services, and agents. Multi-site physical operation remains to be validated.
Standards & Safety: ISO 15118-20, OCPP, and an Enforced Envelope
EV charging is a standards-governed domain. Auralink's source architecture aligns the charger OS to ISO 15118-20 and OCPP and defines a bounded envelope for the agentic layer. This is architectural alignment for inspection and later validation, not certification or validated physical interoperability.
Important boundary: the applicable safety and conformity-assessment requirements depend on the system and intended use, including whether a notified body must be involved. The material on this page is engineering evidence describing how the system is architected for a standards-governed domain — it is not a certification, and it does not assert one. For the engagement that assembles a safety argument, see Hyperion's Safety-Case & Certification.
ISO 15118-20
Road vehicles — Vehicle to grid communication interface — Part 20
ISO 15118-20 specifies the second-generation application and presentation layer for communication between an electric vehicle and the charging infrastructure, including session negotiation, secure communication, and bidirectional power transfer scenarios. It is the standard that governs how a charger and a vehicle establish and run a charging session.
In Auralink
Auralink's charger OS is aligned to ISO 15118-20 at the session layer. For a reference architecture, the relevant point is structural: the system is designed so that vehicle-to-infrastructure negotiation conforms to the governing standard rather than to an ad-hoc protocol.
OCPP
Open Charge Point Protocol
OCPP is the open, vendor-neutral protocol for communication between charge points and a central management system. It is the de facto interoperability standard for EV-charging networks, covering transactions, configuration, firmware management, and diagnostics across heterogeneous hardware.
In Auralink
Auralink's source architecture aligns the charger OS to OCPP and intends the management-plane interfaces to follow a documented open contract. Current protocol interoperability and conformance remain separate validation gates.
How an Edge-Agentic System Is Architected for a Standards-Governed Domain
Autonomy Operates Inside an Enforced Envelope
The agentic layer reasons and proposes within bounded authority. The boundaries — what the system is permitted to command, under which conditions — are defined by the architecture and the governing standards, not by the agents themselves. This separation is the architectural pattern that makes an agentic system defensible in a safety-relevant domain.
Open Standards as the Conformance Substrate
Aligning the charger OS to ISO 15118-20 and OCPP defines a target for externally observable behaviour against documented, auditable contracts. Demonstrated conformance and interoperability require separate validation.
Edge-First for Continuity and Local Determinism
The design targets edge operation and degraded-connectivity continuity, so local decision paths need not require a distant-cloud round-trip. Safe offline behaviour remains a requirement for hardware and field validation.
Sovereign Infrastructure: the Operator Owns the Substrate
The design targets operator-controlled infrastructure and avoids requiring an external cloud in the control path. This sovereignty goal is an architectural requirement, not evidence of an operating deployment.
Evidence
The evidence below is presented honestly and in context. The research reference is a public technical preprint, not peer reviewed; its results are not reproduced here, and it is not a field deployment. The remaining evidence concerns implemented architecture, not private inventory or build speed.
Public technical preprint preprint 2603.08736, "Autonomous Edge-Deployed AI Agents for Electric Vehicle Charging Infrastructure Management" (submitted to arXiv on 24 February 2026). The cited record is a preprint, not evidence of peer-reviewed publication.
Edge-Agent Research — the arXiv Preprint (2603.08736)
The founder's arXiv 2603.08736, "Autonomous Edge-Deployed AI Agents for Electric Vehicle Charging Infrastructure Management" (February 2026), describes edge-deployed agents for autonomous incident resolution. It is a public technical preprint, not peer reviewed; its results are not reproduced here, and it is not field or commercial evidence.
Implemented Architecture Patterns
The reference shows that the patterns described on this page — edge-first decomposition, an agentic layer inside an enforced envelope and standard-aligned interfaces — are implemented in a research system. This does not establish deployment, performance, private repository scale or delivery speed.
A Standards-Governed Domain, by Construction
EV-charging infrastructure is a safety-relevant, standards-governed domain. The relevant evidence is not a single benchmark but the fact that the architecture was built to operate inside that domain's constraints — ISO 15118-20 and OCPP alignment, edge-first operation, and bounded agent authority — rather than retrofitted to them.
What This Is — and What It Is Not
Auralink is a Hyperion-authored Physical AI research reference. On this page it is used as architecture and safety evidence for a standards-governed, safety-relevant domain. It is evaluated in simulation and bounded environments and is documented within a controlled research boundary.
It is not an external client engagement, and no external client names, logos or results are claimed anywhere on this page. The research cited here is arXiv 2603.08736, a public technical preprint, not peer reviewed; its results are not reproduced here. Private inventory is not published.
The preprint is not a refereed publication. And to restate the standards boundary: the applicable safety and conformity-assessment requirements depend on the system and intended use, including whether a notified body must be involved — the material here is engineering evidence, not certification.
Frequently Asked Questions
Is Auralink an external client project?
No. Auralink is Hyperion-authored, documented research. It is presented as an architecture and safety reference evaluated in simulation and bounded environments within a controlled research boundary. It is not an external client engagement, and no external client names or results are claimed here.
What standards apply to Auralink?
Auralink is a research reference for EV-charging infrastructure, a standards-governed domain. Its source architecture is aligned to ISO 15118-20 and OCPP. Current interoperability and physical integration remain separate validation gates.
Can you build this for my safety-critical system?
The relevant Hyperion engagements are Physical AI Deployment (edge-first architecture, agentic layers, and integration for physical systems) and Safety-Case & Certification support (assembling the engineering evidence a safety argument requires). The right starting point is a conversation about your domain, its governing standards, and where an edge-deployed system fits. See /services/physical-ai-deployment and /services/safety-case-certification.
Is the cited research peer-reviewed?
No. arXiv 2603.08736, "Autonomous Edge-Deployed AI Agents for Electric Vehicle Charging Infrastructure Management" (February 2026), is the founder's public technical preprint, not peer reviewed; its results are not reproduced here. It is not a refereed paper, field result or commercial outcome.
Sources and References
Cherifi, M. (2026). "Autonomous Edge-Deployed AI Agents for Electric Vehicle Charging Infrastructure Management." https://arxiv.org/abs/2603.08736
Context: Public technical preprint arXiv 2603.08736 (February 2026), by the founder. Not peer reviewed; its results are not reproduced here. It is not field evidence or a commercial outcome.
ISO (2022). "ISO 15118-20: Road vehicles — Vehicle to grid communication interface — Part 20."
Context: International standard specifying the second-generation application and presentation layer for EV-to-infrastructure communication, including session negotiation, secure communication, and bidirectional power transfer.
Open Charge Alliance (2024). "OCPP — Open Charge Point Protocol."
Context: Open, vendor-neutral protocol for communication between EV charge points and a central management system. The de facto interoperability standard for EV-charging networks.
Discuss Your Safety-Critical System
If you are building Physical AI for a safety-relevant, standards-governed domain — and you want the architecture to stand up to scrutiny — start with a conversation. Hyperion brings research architecture and evaluation experience from Auralink, alongside the founder's production engineering career, and works with your team on deployment decisions and the evidence a safety case requires.
Fractional CPO and Interim Head of Product
Mohammed Cherifi is the founder of Hyperion Consulting and the founder of Auralink, a research reference for an edge-first agentic architecture for EV-charging infrastructure. With long-standing experience in automotive and embedded systems engineering, he specialises in Physical AI deployment for safety-relevant, standards-governed domains.
Related Resources
Physical AI Deployment
Edge-first architecture, agentic layers, and integration for safety-relevant physical systems
Safety-Case & Certification
Assembling engineering evidence for a safety argument; conformity assessment follows the applicable procedure
Sim-to-Real for Industrial Robotics
Closing the sim-to-real gap: physics simulation through on-robot edge inference
Physical AI Product Management
The Product System — Strategy, Discovery, Productization, Launch, Scale