Technical Product Manager

NorthMark Compute & Cloud (NMC²) is backed by dedicated leadership and investment, with a clear mission as it operates at the bleeding edge of technology. Its goal is to scale and enhance the high-performance computing (HPC) and cloud infrastructure that supports its clients' research, production, and delivery, enabling breakthroughs that shape the industries of tomorrow. Its engineers build critical infrastructure to eliminate friction in scientific research, simulations, analysis, and decision-making, accelerating discovery and driving faster innovation.

The Position

NMC² is looking for a Technical Product Manager to own the Network product line within our HPCaaS portfolio. Network is the connective tissue of the platform: every workload, from a tightly-coupled MPI job to a thousand-GPU training run, is shaped by the fabric it runs on. This role exists to make that connective tissue a deliberate product rather than an inherited byproduct of each cluster build.

Our network estate spans the full range from standard Ethernet, through InfiniBand NDR for tightly-coupled HPC and large-scale distributed training, to rack-scale NVLink fabrics — and extends outward to how tenants connect to and beyond the platform, including tenant-defined virtual networks (VPC-style), isolation, cloud interconnect, and peering. As we converge multiple bespoke implementations onto a common platform, the Network product manager will define the fabric tiers we offer, the workload patterns each is built for, and the roadmap for how we evolve them. You will partner closely with network engineering, architecture, and the Compute and Storage product lines on capacity, lifecycle, and what is technically achievable, to ensure the fabric serves the workloads — not the other way around.

Responsibilities

• Product Strategy: Own and evolve the product strategy for Network, developed in close partnership with engineering and grounded in technical reality, ensuring fabric design is driven by customer workload patterns and aligned with the broader HPCaaS portfolio.
• Product Ownership: Own the Network product line end-to-end across Ethernet, InfiniBand, and rack-scale NVLink fabrics. Define what customers can consume, how it is described, and how it interacts with Compute and Storage.
• Fabric Tiering: Define the fabric tiers offered to customers, the workload archetypes each is built for (training, inference, MPI, general-purpose), and the trade-offs between cost, bandwidth, latency, and topology.
• Roadmap Development: Build and maintain a credible roadmap for the fabric, covering near-term scale-out, generational transitions (e.g. NDR to XDR, next-generation NVLink), and convergence across current implementations. Make major platform and vendor decisions collaboratively with engineering and operations, so that significant commitments are sound before they are made.
• Customer Focus: Act as the voice of the HPC and AI engineers whose jobs live or die on the fabric. Translate their communication patterns and scaling pain into clear product decisions.
• Requirement Gathering: Partner with network engineering and architecture to translate workload requirements into fabric specifications, topology choices, and prioritized roadmap items. Make the trade-offs explicit.
• Cross-Functional Leadership: Drive shared understanding across network, platform, and operations teams to shape what gets built, and to deliver and drive adoption of new Network capabilities. Coordinate with the Compute and Storage product managers to ensure end-to-end performance is owned, not assumed.
• Performance Metrics: Define and track the product metrics that matter for Network — sustained bandwidth, tail latency, link health, congestion behaviour at scale, and the workload-level outcomes those drive.
• Financial Transparency (FinOps): Work with finance and engineering to deliver cost transparency on fabric consumption, and inform explicit trade-offs between fabric tier, performance, and price.
• Tenant Connectivity: Own how tenants connect to and beyond the platform — tenant-defined virtual networks (VPC-style), multi-tenant isolation, private interconnect (e.g. cloud direct-connect and cross-cloud peering), and hybrid links back to customer environments. Define the connectivity tiers offered and their place in the product portfolio.
• Risk Management: Identify capacity, performance, security, and operational risks across the fabric, and work with stakeholders to develop mitigation plans — particularly around blast radius and large-scale training resilience.
• Security & Isolation : Make tenant isolation, traffic separation, and encryption in transit explicit product requirements across the fabric and connectivity tiers, partnering with security and engineering on the controls and the compliance posture each tier must meet.

Requirements

• Education: Bachelor's degree or equivalent experience.
• Product Management Experience: Significant product management experience in infrastructure, platform, or deeply technical environments, with a track record of shipping products that engineering customers actually adopt.
• Network Domain Expertise: Deep domain expertise in high-performance networking. Strong working knowledge of InfiniBand (NDR-class fabrics, congestion control, adaptive routing), high-speed Ethernet (RoCEv2, lossless designs), and rack-scale interconnects such as NVLink. Familiarity with Clos and rail-optimized topologies for AI training is strongly preferred. Working knowledge of tenant connectivity — network virtualisation and isolation, private cloud interconnect, and peering — is expected alongside the fabric-side expertise.
• Workload Fluency: Strong understanding of how communication patterns shape fabric design — all-reduce and all-to-all in distributed training, MPI collectives in classic HPC, and the I/O patterns of multi-protocol storage traffic. Able to reason from workload behaviour to fabric requirements and back.
• Platform & Service Thinking: Experience defining or owning network products or service tiers in a multi-tenant infrastructure context, with a focus on consumable abstractions over raw fabric capability.
• Security & Multi-Tenancy : Solid grasp of network security in shared infrastructure — tenant isolation, traffic segmentation, encryption in transit, and the compliance considerations that shape product requirements in a multi-tenant cloud.
• Industry Awareness: Informed perspective on where high-performance networking is heading — generational bandwidth and latency curves, the Ethernet-versus-InfiniBand trajectory for AI fabrics, optical and co-packaged optics developments, and the implications for platform design over a three- to five-year horizon.
• Customer Focus: Strong customer orientation, with empathy for HPC and AI engineering teams and a clear-eyed understanding of how network design choices affect time-to-result and developer productivity.
• Analytical Skills: Excellent analytical and problem-solving skills, with the ability to make data-driven decisions in complex technical environments and to reason about trade-offs from first principles.
• Communication: Outstanding written and verbal communication, with the ability to translate fluently between hardware and platform engineering teams, customer engineers, and senior business stakeholders.
• Influence & Collaboration: Demonstrated ability to lead through influence across engineering, architecture, operations, and commercial teams. Comfortable driving alignment without direct authority.
• Adaptability: Thrives in a fast-paced, evolving environment, with a proactive and pragmatic approach to change.

Location Dallas, TX - Hybrid 3 days onsite

Relocation assistance is available for this opportunity