Role SummaryEvery inference request we serve is a decision about hardware, cost, and performance. Making those decisions well, at scale, across a distributed GPU fleet, is a data problem before it is anything else. The intelligence layer behind those decisions is one of Parasail's real advantages, and it is what this role owns.
Our internal analytics are agent-native. Teams ask the business questions directly and get real analysis in minutes, on data infrastructure designed so that people and agents query it the same way. It is a different way to run a company, and it is a large part of why a small team moves as fast as we do.
You will own and expand that layer. You will design the data models behind fleet and unit economics, build the analytical capabilities the rest of the company runs on, and turn questions that used to take a week into answers that take a minute. Work you ship compounds, because it becomes something the entire team can use immediately.
You will work through MCP connections and agentic tools every day. We expect you to be productive in that stack from week one.
Questions You Will Answer- What does a token actually cost us across hardware generations and workload profiles, and where does that curve bend?
- Which workloads run best on which hardware, and what is that difference worth to a customer?
- Where is capacity earning, and where is it waiting, across every GPU and every region we operate in?
- How does a model's benchmark performance translate into real unit economics at production scale?
- As the fleet gets more efficient, how much of that can we hand back to customers as better pricing?
Key Responsibilities- Design and own the data models behind fleet economics, capacity, and utilization
- Build the analytical capabilities and reporting the company makes decisions from, and expand what our internal agentic tooling can answer
- Build and maintain the pipelines that bring infrastructure, usage, billing, and customer data into one coherent view
- Develop dashboards and models connecting operational performance to revenue, margin, and capital efficiency
- Build verification into the data layer so answers stay trustworthy as the product evolves quickly
- Partner with infrastructure, finance, product, and go-to-market to anticipate what they need to know
- Use agentic tooling to automate analysis and reporting, and raise the ceiling on what one person can cover
Required Qualifications- 2 to 5 years in business intelligence, data analytics, analytics engineering, operations analytics, or a similar hands-on data role
- Strong SQL, plus Python or a comparable language for data work
- Daily proactive use of agentic AI tools with MCP connections to pull data from source systems
- Real curiosity about how a business works underneath the metrics, and the drive to chase a number until you understand it completely
- Technically literate enough to work directly with infrastructure and engineering teams and reason about what the systems are doing
- Ability to explain what the data says, and how confident you are in it, to any audience
- High ownership and strong instincts for what is worth measuring
- Excited by a fast-moving startup where you get to define your own scope
Nice to Have- Experience in operations, manufacturing, logistics, or another environment where asset utilization and yield drive the business
- Cloud infrastructure, data center, GPU, or semiconductor exposure
- Usage-based or consumption-based billing experience
- Experience building internal tools, pipelines, or automation
- Familiarity with dbt, warehouse modeling, or observability tooling
Compensation & Benefits- Competitive base salary with meaningful equity
- Comprehensive health, dental, and vision coverage
- Generous PTO and company holidays
- 401(k) and family-planning benefits
- High visibility with the executive team and direct exposure to how the business is run
- The chance to build agentic data infrastructure that the entire company runs on
