Director of Infrastructure Operations

The Director of Infrastructure Operantions will own GPU capacity planning, compute economics, cost engineering, and long-range infrastructure strategy - sitting at the intersection of Finance and Engineering. This person will build the end-to-end economic model of compute, drive infrastructure investment decisions, and manage relationships and proposals across the infrastructure supply chain., * Lead multi-year GPU capacity planning and demand forecasting

• Build fully loaded cost models and unit-economics metrics for compute infrastructure
• Evaluate buy/lease/cloud infrastructure decisions and build supporting business cases
• Translate capacity needs into CapEx/OpEx and cost-of-goods impacts
• Partner with Finance on operating plans and long-range forecasts
• Build models and dashboards for capacity, utilization, and cost visibility
• Serve as the key liaison between Engineering and Finance on infrastructure economics

• 10+ years in data-center economics, capacity planning, or cost engineering
• Background in large-scale compute, cloud, AI/ML, GPU, semiconductor, or data-center environments
• Strong grasp of unit economics, CapEx/COGS modeling, and depreciation
• Advanced Excel/Google Sheets modeling skills
• Bachelor’s degree in Economics, Engineering, Computer Science, Operations Research, or a related field., * MBA or Master’s degree in a technical or quantitative field
• Experience at a GPU cloud provider, hyperscaler, or semiconductor company
• Familiarity with GPU clusters, networking, and storage infrastructure

• A pivotal role shaping Cognichip’s compute strategy and infrastructure economics from the US HQ
• Direct partnership with Finance, Engineering, and business leadership
• Work at the intersection of AI, semiconductors, and infrastructure, solving high-impact economic and technical problems
• Competitive compensation package, including early-stage equity in a fast-growing AI startup
• A high-trust, low-ego team culture that values ownership, curiosity, and collaboration

Cognichip