Quantitative Researcher
About Auriko
Auriko is building the market infrastructure for AI inference.
Our current product is an inference platform built with quantitative trading rigor. It enables AI teams to quickly switch models across inference providers and optimize inference cost.
Our team comes from quantitative trading and high-frequency derivatives backgrounds, with experience building, trading, and risk-managing complex systems at scale.
Role
You will research, design, and validate the quantitative models that drive the platform's routing, pricing, and risk decisions.
You would be responsible for:
- Routing optimization: research and design constrained optimization methods to route inference requests across providers (filter → score → select), subject to user constraints (latency, cost, uptime) and platform constraints (capacity, rate limits), consistent with portfolio allocation principles
- Pricing and cost models: develop quantitative pricing and relative-value models to evaluate provider economics, quantify cost under varying conditions, and model expected savings and margin impact across routing alternatives
- Predictive signals and risk: research, design, and validate predictive performance signals using time-series and statistical methods—forecasting latency distributions, provider health, and performance changes that feed real-time routing decisions
- Backtesting and simulation: design frameworks to evaluate routing strategies under stress scenarios including provider degradation, price changes, capacity shocks, and partial outages; conduct sensitivity analyses to validate robustness
- Data specification and metrics: define model specifications, data requirements, and statistical quality control for platform measurements; design analytics pipelines for cost breakdown, performance attribution, and model validation
Requirements
- Quantitative research experience: statistics, optimization, time-series analysis, or mathematical modeling
- Numerical and scientific computing: implementing predictive models, machine learning pipelines, and simulations
- Experience designing and validating predictive models against real-world data
- Comfort with constrained optimization: you can formulate a problem, solve it, and reason about sensitivity and edge cases
Strong Pluses
- Background in quantitative finance, derivative pricing, or portfolio optimization
- Experience applying financial modeling principles (relative-value, risk management, allocation) to non-financial domains
- Monte Carlo simulation, distribution modeling, or stochastic processes
We are building a small, high-ownership team. We value intellectual honesty, independent thinking, and fast execution. We care about substance over appearance.
We look for driven, high-agency people and incentivize them to do exceptional work. Our compensation package is market-leading, with meaningful equity upside.
Auriko sponsors work visa.
Contact
join@auriko.ai