nVent - Data Solutions

01 The One-nVent Vision

My nVent journey began with EdgeNRG, an integrated edge data center that leaned on nearly every division under one roof at nVent. nVent is a global leader in electrical connection and protection, and while most people have never heard the name, almost everyone in the electrical world knows its brands: ERICO, HOFFMAN, SCHROFF, ERIFLEX, CADDY, RAYCHEM, ILSCO, and TRACHTE^[1], extended through acquisitions such as CIS Global (intelligent rack power distribution), Eldon (enclosures), ECM Industries (electrical consumables), and TEXA (thermal management). That breadth (power distribution, connectors, enclosures, racks, cabling, and cooling all in one company) is exactly what made EdgeNRG possible.

EdgeNRG was born from a simple but powerful observation by my former manager, David Wood. If nVent already makes the power distribution, the battery backup and redundant power, the rack, and the rack cooling, why not become the integrator and combine all of it into a single, decentralized edge data center node? A small, self-contained data center you could drop anywhere the data is generated.

EdgeNRG became integrated physical infrastructure for edge nodes: rugged cabinets, power management, and remote management tools, with DC or AC power and optional UPS battery backup, built for 5G and 6G edge computing and reaching into smart cities and intelligent infrastructure (a space I represented nVent at conferences beginning in 2023). One of the prototype’s headline features, and the thread that would define my early years here, was switched and metered DC power distribution^[2].

02 Driving the Smart DCDU: Requirements to Loading Dock in 87 Days

A telecom customer in Finland saw the switched and metered DC distribution on the EdgeNRG prototype and recognized it as the replacement for an obsoleted product they depended on. I captured their requirements, walked their site, and built a complete picture of both their technical needs and their real-world deployment environment. Then two of us, a genuine skunkworks team, carried the product from a blank requirements document all the way to shipping hardware on their loading dock in 87 days. I drove the architecture, coordinated the global team, and owned the outcome.

This was a true One-nVent product, assembled from the strengths of silos that rarely build together:

• Chassis and card cage from nVent Schroff
• Custom printed circuit board assemblies from nVent Wuxi
• Power distribution blocks from nVent Erico
• Terminal blocks and compression lugs from nVent ILSCO
• DIN rail from nVent Hoffman
• Controller, firmware, software, and cybersecurity from nVent SmartPower

The custom PCBAs are what make 87 days genuinely wild to anyone who has shipped hardware. Walk a perfect schedule: two weeks for schematics, two weeks for layout, three weeks for fabrication, one week for assembly, two weeks for shipping. That is ten weeks, or 70 days, with everything going right and nothing left over for testing, integration, bring-up, firmware and software, or the product itself. We hit the window because the boards came back correct on the very first pass, and because there was no slack anywhere in the plan.

To run a schedule that tight, I worked around the clock and followed the sun. Early mornings I synced with our firmware team in Bangalore. During the day I worked my own hours in New Jersey and stayed late to support Arizona’s end of day. Around 10 PM I took calls with Wuxi, China on board fabrication and design for manufacturability. Around 1 or 2 AM, as Straubenhardt, Germany came online, I coordinated the chassis, card rails, and custom faceplates. The plans had to be perfectly synchronized. We could not afford to wake up at 8 AM in New Jersey and discover that Germany had lost an entire workday because a question went unanswered overnight.

With a team of two, I wore every hat the schedule demanded: the PCB and its design for manufacturability, the supply chain and expediting, the chassis tolerances, the mechanical, layout, and schematic reviews, the system architecture, the firmware and software integration and bring-up, the throwaway prototypes the software teams coded against until real hardware arrived, the UX and new features, shipping, and compliance. Each of those is its own discipline, with its own people to coordinate across time zones, and getting the product out the door meant carrying all of them at once. The customer had never seen a requirements document become finished hardware on their dock that fast, or that customized. They chose nVent for their long-term infrastructure, and we hardened the design iteratively into a strong standard product: the nVent Smart DCDU shipping today^[3], delivering intelligently managed -48V DC power for mission-critical telecom and edge environments across public safety, utilities, transport, and defense.

03 Certification, Deployment, and Customer Support

Once the design was proven, I carried the Smart DCDU through UL, IEC, CSA, and ETL certification for both product safety and EMI/EMC compliance, spanning standards such as UL 62368-1, ETSI EN 300 386, and EN 61326-1 (Class A). As the product owner through this phase I captured the requirements, mapped which standards applied, and worked directly with the test houses to agree on which tests were relevant and how they would run. Compliance is not a paperwork exercise you delegate. I traveled to Intertek’s facility in Lake Forest, California to set up, witness, and guide the Safety and EMI/EMC qualification in person. Being on the floor while the equipment is under test, answering questions as they come up, is how you catch problems early and keep a launch on schedule.

Beyond certification, I have spent these years supporting customers directly through commissioning, deployment, integration, and troubleshooting, and channeling what I learn on the ground back to our architecture and engineering teams as concrete feedback to improve the product line. Once the Smart DCDU launched, I pivoted into a broader R&D role within nVent Data Solutions, chasing opportunities across the SmartPower and Liquid Cooling verticals and their integration into the larger nVent portfolio.

04 Reconnaissance and Market Intelligence

Across my time here I have served as nVent’s lead for reconnaissance and market intelligence on the floors of trade shows and conferences (OCP, GTC, SC, DCD, DCW, OCP EMEA), as well as on customer site visits and in Voice of Customer engagements. I survey the mechanical, electrical, software, and system-architecture designs on display and translate them into competitive-landscape and market-positioning intelligence for our R&D leadership.

What makes this valuable is perspective. My background spans Electrical, Computer, Robotics, and Software Engineering across space, automotive, autonomous vehicles, defense, and now data centers. That multidisciplinary lens lets me read a competitor’s design intent and spot where the market is heading. I also scope how competitors implement compliance and patents on their products, and how they navigate the regulatory and certification landscape to get there, then bring that intelligence back to the team with technical depth rather than just a photo of a booth.

05 Remote-First Networking and Connectivity

EdgeNRG’s whole premise, decentralized nodes deployed wherever the data is, depends on remote and reliable access to hardware. And I have been fully remote the entire time, which meant I was the end user of my own decisions and lived with all of their pros and cons. Developing on the same remotely connected and serviced hardware we were shipping surfaced the real gaps in our remote testing and testability. Failing fast and learning fast, I started digging into cellular modems, redundant network systems, fail-safe engineering practices, and overlapping cellular coverage maps, and I brought Starlink into the picture as its coverage expanded.

That interest never left. At home I run a 25U Eaton Wright Line rack packed with equipment: network cable cards, network video recorders, switches, modems and routers, servers running Docker, plugins, virtual machines, Unraid and NAS, UPS battery backup, and local AI nodes running models on-prem. It is a working lab that keeps my networking, virtualization, and infrastructure skills sharp. You can see more of that build on my Server and homelab page.

06 Cybersecurity and Product Security

My combined hardware, software, and networking footprint in Data Solutions drew me into nVent’s Product Security Advisory Team (PSAT), deepening a cybersecurity focus that began with my academic work at WPI. That work spans system diagramming, risk assessment, and threat modeling, making sure the connected power and cooling products we ship are secure by design, not just functional.

07 Technical Lead for HVDC

High-voltage DC was a logical progression from the low-voltage DC PDU work behind the Smart DCDU. Building on the deep DC familiarity I earned shipping that product, I took on Technical Lead for nVent’s HVDC portfolio on top of my existing role.

High-voltage DC is where AI-factory-scale power is heading, and I drive the systems, architecture, and engineering direction for nVent’s emerging HVDC work, assessing how the DC power path could run from facility utility down toward GPU compute. The field is still in its infancy, with few established players, so nVent’s work today is integration and test rather than ground-up development, and I lead that effort across the two architectures the industry is converging on.

NVIDIA is pursuing an 800 VDC bus, while much of the rest of the field favors ±400 VDC. That difference matters most for touch-safety: a ±400 VDC rail keeps a technician no more than 400 V above ground on contact with either pole, whereas an 800 VDC bus can leave that technician 800 V from ground at the same delivered potential. The ±400 VDC camp is far from monolithic, though: the hyperscalers each bring their own flavor, from the Open Compute Mt. Diablo specification to different grounding schemes and floating versus non-floating references. Part of my job is tracking that fragmentation as the standards settle, while NVIDIA works with UL and the certification and regulatory bodies to qualify its 800 VDC path.

I hold regular calls with NVIDIA and AMD on their evolving power architectures, evaluate HVDC power supplies and power shelves, and pay close attention to serviceability and to how the HVDC sidecar interacts with the IT compute rack. To stay ahead of where the standards are moving, I attend the Open Compute Project’s working calls and reach out directly to customers, suppliers, and vendors on emerging requirements.

08 AI-Augmented Engineering and AI Leadership

In 2026 I leaned hard into the new generation of AI development tools: Codex, Grok Build, Claude Code, and others. Using them, I began designing, building, and shipping real software: applications, features, configurators, testing scripts, and automations. They let me build and deploy the kinds of software products I had always wanted to make.

That work grew into a broader role. I am now one of nVent’s AI leaders, tasked with expanding the use and adoption of AI across the company. I host and run workshops and sessions with our software engineering teams on how to use, leverage, and expand AI capability, and I continuously benchmark, assess, and test the models as they release, weighing cost against capability and reporting back to software management so we can roll the strongest options out to the broader team.

The clearest example is the configurator I built for nVent’s modular Technology Cooling System (TCS). It runs algorithms that solve for the optimal configuration of an entire liquid-cooled system: the fitment and routing of pipes, the distances to equipment, hose and manifold selection, pipe rotations, elevations, and clearances, then generates the bill of materials for a shippable, deployable design. A tool like that would normally be an expensive, outsourced software engagement. With AI tools, I developed it internally, on my own. It keys directly into the deployment problem: turning a cooling layout into a quote and a deployable system, fast. nVent’s public AI Deployment Center^[4] and its online configurator showcase the direction of this modular, deploy-ready approach to liquid cooling and power.

09 The Siemens and nVent 100 MW AI Factory Blueprint

This is the clearest expression of where my work points: the power and cooling equipment that gets deployed inside an AI factory. I helped develop the Siemens and nVent 100 MW Hyperscale AI Blueprint^[5], a Tier III reference architecture that integrates Siemens industrial-grade electrical systems, nVent liquid cooling, and NVIDIA GB200 NVL72 systems (127 kW per rack) into a modular, 100 MW AI data center. Working within a joint Siemens and nVent team, I contributed the power-distribution and liquid-cooling content: the drawings, diagrams, and technical detail behind the published design.

10 References