Director, Presales Solution Architecture - NeoCloud

Important: if an employer asks you to log into their system via iCloud or Google, send a code, an SMS or Telegram password, run some code, or install software — refuse. These are signs of fraud.

, an IREN company, is the Kubernetes-native AI infrastructure company, enabling organizations to build and operate scalable, secure, and sovereign infrastructure for modern AI, machine learning, and data-intensive applications. By combining open source innovation with deep expertise in Kubernetes orchestration, empowers platform engineering teams to deliver composable, production-ready developer platforms across any environment—on-premises, in the cloud, at the edge, or in sovereign data centers.

As enterprises navigate the growing complexity of AI-driven workloads, delivers the automation, GPU orchestration, and policy-driven control needed to manage infrastructure with confidence and agility. Committed to open standards and freedom from lock-in, ensures that customers retain full control of their infrastructure strategy.

Job Description

Why this role exists

K0rdent AI is the orchestration layer that turns raw, disaggregated GPU infrastructure into a multi-tenant, production-ready AI cloud — without locking companies into a single hyperscaler or hardware vendor. We sell accelerated compute: GPU clusters, bare metal, and managed AI infrastructure to Neoclouds, AI-native startups, enterprise AI teams, research labs, and sovereign/regulated buyers. These are technical, high-value, long-cycle deals where the sale is won or lost on credibility: whether we can architect the right cluster, model the real TCO, prove performance, and de-risk a customer's move onto our platform.

This person owns the technical win. They build and lead the sales engineering function that turns "interested" into signed, multi-year committed-capacity contracts, and they set the pre-sales bar as we scale headcount and deal volume.

This is not a demo-jockey role. We need someone who has genuinely stood up training and inference workloads, argued interconnect topology with a customer's ML infra lead, and closed large deals with cycles measured in quarters, not weeks.

What you’ll own

Lead and build the SE / Solutions Architect team

Hire, coach, and retain a team of sales engineers and solutions architects; define the pre-sales operating model as the org scales.

Build the reusable machinery: discovery frameworks, reference architectures, TCO/benchmark models, POV playbooks, demo and benchmark environments, RFP response libraries.

Set and hold a technical quality bar across the team; run enablement so every SE can speak credibly to GPU architecture, networking, and orchestration.

Own the technical win in large, complex deals

Partner with Account Executives as the technical lead on strategic and enterprise opportunities from discovery through technical close.

Run qualification with a real methodology (MEDDPICC or equivalent) — surface the economic buyer, decision criteria, and the technical champion, and build the win plan around them.

Architect solutions across compute, networking, storage, and orchestration; produce sizing, capacity plans, and TCO comparisons vs. hyperscalers and self-build.

Design and drive POCs/POVs: define success criteria up front, run benchmarks, and convert results into commercial momentum.

Be the Technical voice of the Customer internally

Feed structured product and capacity requirements back to product, platform, and supply/capacity planning.

Work alongside the NVIDIA field and partner ecosystem (Cloud Partner program, reference architectures, joint pursuits) to strengthen deals.

Influence roadmap and packaging based on what you learn in the field.

Qualifications

Required:

You have actually run or stood up ML workloads — distributed training and/or production inference — not just talked about them.

Practical fluency in the training and inference lifecycle: data pipelines, distributed training (multi-node/multi-GPU), fine-tuning, and serving; you understand where bottlenecks actually live (interconnect, memory bandwidth, I/O, scheduling).

Comfortable in the frameworks and tooling customers use — PyTorch and the surrounding ecosystem…