Senior Technical Presales Engineer, AI & HPC Systems

Semidynamics; Barcelona; On-site; EUR 90,000 - 130,000; Full time

Job summary

Semidynamics, a Barcelona-based fabless semiconductor company, seeks a Technical Presales Engineer to own the technical side of tenders, RFIs and RFPs. Translate architecture into credible, compliant responses: configurations, performance projections and benchmarks, collaborating with BD, CTO and engineering in Barcelona.

You will design customer configurations, define benchmark strategies (MLPerf, LLM metrics, HPC tests) and build reusable collateral.

Qualifications

• 8+ years in technical presales or HPC/AI performance roles with customer interaction.
• Background in server OEM, system integrator or chip maker in AI/HPC space preferred.
• Experience with public HPC procurements or EuroHPC is a plus.

Responsibilities

• Own the technical response to tenders, RFPs and RFIs, including requirement tables and compliance matrices.
• Read documents quickly, flag gaps, and drive internal answers to deadlines.
• Provide inputs for offers: system specs, performance per node, power/thermal figures, SLAs.
• Build and maintain a reusable technical collateral library.
• Define and run benchmark strategies: MLPerf Inference, LLM metrics, HPC benchmarks.

Skills

Technical presales

Solutions architecture

HPC/AI performance

Customer facing

English proficiency

Education

Degree in Computer Science or related field

Tools

Linux

PyTorch/ONNX

Slurm

Kubernetes

PCIe/RDMA

Job description

Europe - based in Barcelona

Solutions Architecture · Tenders & RFPs · Benchmarking · Senior Individual Contributor

Travel: around 25%, mainly to customers, OEM partners and HPC events in Europe

Semidynamics is a Barcelona-based fabless semiconductor company developing a complete silicon solution for AI inference and HPC in the cloud. Built on RISC-V and a disruptive memory architecture, our accelerators are purpose-built for LLMs, Agentic AI, and general compute. Our systems, from PCIe cards and servers to full liquid-cooled racks, are being offered into European AI Factories, AI Gigafactories, HPC centres and cloud operators, usually together with OEM and system integrator partners.

THE OPPORTUNITY

Our commercial pipeline is growing faster than our ability to answer the technical questions it generates. European HPC and AI infrastructure is bought through formal tenders: every opportunity comes with requirement tables, benchmark rules, power and cooling questions, compliance annexes and partner documentation, often on several bids at once and on tight deadlines.

As our first Technical Presales Engineer, you will own the technical side of the sales process. You will be the person who turns our architecture into credible, compliant answers: system configurations, performance projections, benchmark results and technical responses that win tenders and convince technical evaluation teams. You will work with the BD team, our CTO and architecture, performance and software engineering in Barcelona.

WHAT YOU WILL DO

Tenders, RFPs and RFIs

• Own the technical response to tenders, RFPs and RFIs, both direct and as a subcontractor inside OEM or consortium bids: requirement tables, compliance matrices, technical annexes, risk and roadmap sections.


• Read tender documents quickly, extract what applies to Semidynamics, flag gaps and drive internal answers to deadline.


• Supply OEM and integrator partners with the inputs they need for their offers: system specifications, performance per node, power and thermal figures, support and SLA terms, and technical sections ready to paste into their responses.


• Build and maintain a reusable answer library and standard technical collateral.

Benchmarking and performance

• Define and run the benchmark strategy for opportunities: MLPerf Inference, LLM serving metrics (throughput, time-to-first-token, latency per token), HPC benchmarks (HPL, HPCG, application kernels) and tender‑specific scoring methods.


• Produce and sanity‑check performance projections and roofline analyses with our performance and software teams; make sure every number we publish is defensible, consistent and compared against a current competitor baseline.


• Prepare TCO, performance‑per‑watt and cost‑per‑token comparisons against GPU and accelerator alternatives.

Solution architecture

• Design customer configurations from card to rack: host CPU choice, PCIe topology and lane budget, accelerator‑to‑host ratio, memory, networking (Ethernet/RoCE, InfiniBand, Ultra Ethernet), storage, rack power, liquid cooling and facility requirements.


• Work with OEM/ODM partners and the server platforms we integrate into (19-inch and OCP ORv3) on integration and qualification questions.


• Help build bills of materials and configuration inputs for pricing, together with BD and Operations.

Customer technical engagement

• Be the technical counterpart for HPC centre and data‑centre technical teams: architecture deep‑dives, workshops, Q&A on data types, precision, software stack, programming model, multi‑tenancy, security and code portability.


• Deliver technical presentations to customers, partners and evaluation committees.


• Scope and support evaluation programmes, early‑access systems and joint R&D or benchmarking collaborations.


• Feed customer requirements back to Product and Engineering in a structured way.

CANDIDATE PROFILE

• Experience: 8+ years in technical presales, solutions architecture, systems engineering or HPC/AI performance roles, with a significant part customer-facing.


• Background: experience at a server OEM, system integrator or chip maker in the AI/HPC space (e.g. HPE, Dell, Lenovo, Eviden, Supermicro, Gigabyte, NVIDIA, AMD, Intel, or comparable) is strongly preferred.


• Tenders: has written technical responses for public HPC or data‑centre procurements; experience with EuroHPC or national supercomputing tenders is a strong plus.


• Hands‑on: has configured, benchmarked and tuned AI or HPC systems, not only described them.


• Communication: writes clear, precise technical English under deadline pressure; equally comfortable with engineers, procurement teams and executives.


• Languages: fluent English required; Spanish, French or German a strong plus.


• Education: degree in Computer Science, Electrical/Computer Engineering, Physics or a related field; MSc or PhD advantageous.

SPECIFIC KNOWLEDGE REQUIRED

• HPC and AI server architecture: CPUs (x86, Arm, RISC‑V), accelerators, PCIe Gen5/Gen6 and switches, CXL, memory hierarchy (HBM, LPDDR, DDR) and bandwidth limits.


• Rack‑scale design: node and rack power budgets, direct liquid cooling and CDUs, OCP vs 19‑inch, data‑centre facility constraints.


• Cluster networking and RDMA for AI and HPC (Ethernet/RoCE, InfiniBand, Ultra Ethernet).


• LLM inference: prefill vs decode, batching, KV cache, quantisation and number formats (FP4/FP8/BF16/FP64), serving frameworks.


• Benchmark methodology: MLPerf rules, HPL/HPCG, how to build and defend projections.


• AI and HPC software stack: Linux, PyTorch/ONNX, BLAS and math libraries, compilers, containers, schedulers (Slurm, Kubernetes).


• Familiarity with public procurement mechanics (requirement tables, scoring, compliance annexes, EU‑content rules) is a plus.

WHAT WE OFFER

• The founding presales role in a European AI silicon company, with real influence on how products are positioned and configured.


• Close work with world‑class architecture and performance engineers in Barcelona.


• Exposure to the largest AI and HPC infrastructure projects in Europe.


• Competitive compensation aligned with our job architecture and salary bands.

Semidynamics is an equal‑opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.