What you'll do
- Extend the solver core — reachability, viability, contagion, causal inference, narrative generation, geospatial, time-series, and domain-specific engines. Proprietary. Closed-source.
- Ship new decision domains. A domain pack is a declarative spec of a state/action/disturbance system plus a calibration harness. Typical domain = one month of work, one engineer, one domain expert.
- Own calibration — the Brier-score discipline that turns forecasts into credible verdicts. Everything we ship is graded against outcomes.
- Build the platform that makes the engines usable by non-engineers: NL parsing, explanations, visualizations, SDKs.
Who we hire
Senior engineers with a production track record in one or more of: mathematical optimisation, probabilistic programming, causal inference, agent-based simulation, real-time signal processing, or enterprise decision software. We value deep depth over broad breadth.
We do not run technical-interview brain-teasers. Our interview is a half-day working session against a realistic problem from the codebase, paired with one of us. We pay for your time.
How we work
- Remote-first, headquartered in Toronto, Canada. Core hours are anchored to Toronto / ET; some roles require CET or PT overlap depending on customer footprint.
- Small team. No engineering-manager career ladder until the company is meaningfully larger. You will ship code, talk to customers, and debug production.
- Closed-source, always. Your work stays inside the company. If you need to publish, we have a formal process for gated methodology whitepapers that expose results without exposing implementation.
- Meaningful equity. Early-stage compensation bands trade cash for ownership. If that is not what you're optimising for, ARIAA is the wrong fit.
Current openings
We do not maintain a rolling list of roles; we open searches when a domain requires it. If you match the profile, email us with your strongest production artefact and a one-paragraph note on what you would build first at ARIAA.