SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Silicon Data builds data infrastructure for the global compute economy, providing price transparency, benchmarking, and market indices for AI infrastructure participants. The company is backed by leading trading firms (DRW, Jump Trading) and recently partnered with CME Group to launch compute futures markets.
SiliconMark is Silicon Data's measurement layer for transacting compute—a third-party GPU performance benchmark suite that addresses a critical market gap: actual GPU performance can vary by up to 34% for the same hardware spec, creating friction for compute buyers and sellers. This role owns SiliconMark end-to-end: the benchmark suite, agent, scoring methodology, and certification layer that customers use as evidence in commercial negotiations.
You will define and execute the SiliconMark roadmap, including benchmark coverage, agent capabilities, cohort scoring, and the self-service surface for running and interpreting jobs. Key responsibilities include: deciding which hardware and application benchmarks belong in the suite and how they are scoped, versioned, and deprecated; partnering with GPU benchmarking engineers on methodology (cohort definition, statistical thresholds, run repeatability, configuration fingerprinting); conducting discovery with infrastructure teams, AI/ML teams, procurement teams, and resellers to understand workflows and evidence requirements; translating customer needs into specs and prioritized backlogs for Engineering; driving go-to-market planning for SiliconMark launches with distribution partners (Bloomberg, Refinitiv, direct enterprise); defining key product metrics (devices benchmarked, cohort population, run success rates, time-to-first-result, trial-to-paid conversion); connecting SiliconMark to wider pricing products to express performance in economic terms; and communicating product direction through documentation, methodology write-ups, and public materials.
This is an early-stage, high-ownership role at the intersection of data products, distributed systems, and trust products. You will work closely with the Founder/CEO, Engineering, and domain-expert GPU benchmarking engineers. The team values ambition, comfort with ambiguity, and strong opinions about enterprise data products.
REQUIREMENTS:
- 5–7 years of product management experience related to GPU hardware
- Demonstrated ability to own products end-to-end (problem definition through delivery and iteration)
- Comfort working with complex or structured data products
- Strong analytical instincts
- Technical fluency with infrastructure: credible conversations about GPU cluster architecture, performance metrics (sustained vs. peak FLOPs, cross-node bandwidth, throughput, latency), drivers, firmware, and cloud deployment models
- Excellent written communication (clear PRDs, crisp specs, concise stakeholder updates; ability to explain methodology decisions to skeptical customers)
- Experience working with enterprise customers (procurement, integration requirements, long-term account relationships)
- Entrepreneurial mindset and comfort in environments where roadmap is still being written
NICE TO HAVE:
- Hands-on exposure to compute benchmarking, stress testing, or performance analysis
- Familiarity with GPU telemetry via driver APIs or benchmarking tooling for GPU clusters and tensor workloads
- Understanding of GPU hardware economics, cloud compute pricing, or AI infrastructure procurement
- Experience with products where an agent or SDK runs inside the customer's environment (deployment, versioning, trust problems)
- Background in certification, attestation, provenance, or audit-grade data products
- Background in or familiarity with financial data products
- Prior experience at early-stage startup building PM function from the ground up
