Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

REW Technology

NewBe an early applicant

Senior Compute Architect

Posted

Description

Job Overview
We design and deliver the physical and logical compute foundation for some of the most demanding environments in the industry — from large-scale enterprise estates to GPU-dense AI factories. The Compute Architect owns the server and accelerator layer end to end: how it is specified, standardised, designed, built, validated, and handed over so that it performs predictably at scale.
This is a hands-on architecture role. You will write the HLDs and LLDs, run design workshops with customers and delivery teams, define the build standards everyone else follows, and be the person the escalations land on when something does not behave the way the paper says it should.
You will sit at the seam between compute, network, storage, and facilities — the layer where most complex delivery failures actually originate. We need someone comfortable owning that intersection rather than defending a single silo.

Key responsibilities
Architecture and design
  • Own the High-Level Design and Low-Level Design for server and accelerator platforms, from rack elevation and cabling topology through firmware baselining and OS/hypervisor layout.
  • Translate customer requirements and workload profiles into a validated, supportable bill of materials and reference design.
  • Produce architecture decision records and design documentation of a quality that delivery engineers, support teams, and the customer can all work from without further interpretation.
  • Ensure every design is compliant with vendor support matrices, security baselines, and our internal standards — and can be proven to be compliant when it is audited.
Standardisation and engineering rigour
  • Build, maintain, and govern the compute build standards and platform catalogue: firmware/BIOS baselines, driver and OS version pinning, PCIe and NUMA layout rules, and NIC/DPU integration patterns.
  • Define the validation and acceptance criteria for each reference design — power-on, burn-in, performance baseline, firmware drift checks, and failure/failover testing — and sign off platform readiness.
  • Own the technical relationship with OEM engineering (Dell, HPE, Lenovo) and silicon/accelerator vendors, including escalation of firmware and integration defects.
Delivery and troubleshooting
  • Lead design workshops and design reviews with customers, delivery, and operations teams; drive open issues to technical closure.
  • Act as an escalation point for cross-domain faults spanning compute, network, storage, and facilities — power and thermal constraints, PCIe and riser topology issues, NIC/DPU and RoCE/InfiniBand integration, firmware interactions, and performance anomalies at rack scale.
  • Support build and handover: validate as-built versus as-designed, resolve drift, and contribute to runbooks and operational documentation.
Leadership and enablement
  • Provide technical direction and mentoring to delivery and operations engineers; raise the floor on compute depth across the wider team.
  • Contribute to pre-sales and bid work: solution shaping, technical responses, sizing and effort estimates, and customer technical presentations.
  • Feed field learnings back into standards, reference architectures, and product direction.
Must-have requirements
  • Deep server and compute architecture expertise — a real working command of CPU and memory architecture (NUMA, memory modes, DIMM population and channel rules), PCIe topology and lane allocation, firmware/BIOS behaviour and tuning, and NIC/DPU integration.
  • Large-scale OEM infrastructure — proven design and deployment experience on Dell, HPE, and Lenovo server platforms at significant scale (multi-rack, multi-hundred to multi-thousand node estates). Familiarity with their management toolchains (iDRAC/OME, iLO/OneView, XClarity) and their support and firmware-release models.
  • Linux administration and troubleshooting at depth — comfortable in kernel and userspace diagnosis: dmesg, journald, lspci/lsscsi/numactl, /sys and /proc inspection, hardware error and MCE/AER analysis, driver and firmware version management, and performance profiling. You can debug a problem rather than follow a runbook for it.
  • GPU and accelerator platforms — strong understanding of GPU and high-density compute architecture: GPU topology and NVLink/NVSwitch, PCIe versus SXM form factors, power and thermal envelopes, MIG, GPU-to-fabric integration, and the operational realities of air-cooled and liquid-cooled high-density racks.
  • Design authority — demonstrated ownership of HLD/LLD, running design workshops and design reviews, authoring build standards, and defining and executing validation and acceptance plans.
  • Cross-domain troubleshooting — a track record of diagnosing issues that cross compute, network, storage, and facilities boundaries, and of driving them to root cause with vendors rather than to workaround.
  • Stakeholder capability — able to hold a technical design conversation with a customer architect in the morning and an RFI response in the afternoon, and to defend a design decision with evidence.
Nice to have
  • NVIDIA HGX / DGX platforms — including current-generation B200 / B300 (Blackwell) systems and their power, cooling, and fabric requirements.
  • AI Factory or HPC delivery experience — cluster-scale build-out, scheduler and fabric integration, benchmarking, and workload validation.
  • NVIDIA reference architectures and NAIC experience — designing against and adapting vendor reference designs rather than starting from a blank page.
  • Liquid cooling and high-density rack integration — CDU, rear-door heat exchangers, and direct-to-chip cooling.
  • Scripting and automation — Python, Ansible, or Bash for build and validation, plus infrastructure-as-code and lifecycle management tooling.
  • AI platform software awareness — Kubernetes, GPU Operator, or NVIDIA AI Enterprise, and familiarity with how the compute layer is consumed by AI platform software.
  • Relevant certifications — Dell, HPE, Lenovo, NVIDIA, or Linux/RHCE, and ITIL-adjacent service transition experience.
Our Culture :
  • Open communication and a focus on psychological safety
  • Recognition of achievements and contributions
  • Collaboration across multicultural teams
  • Less bureaucracy, more focus on results
Hiring Process:
  1. Technical Interview
  2. Managerial Interview
  3. Client Interview
  4. Pre-offer stage + Reference Check (if requested)
  5. Official Offer

Skills

What Senior Architecture jobs ask for — and how much of it you have →
Apply

See also

Architecture jobs by country — openings, pay and top skills →