Senior SDET

Algo•Calgary, AB

About The Position

Software is entering a period where code generation is accelerating, and verification is the new constraint. We are rebuilding our quality function from the ground up as an AI-first verification engineering team, and we are hiring founding SDETs to design and build it. This is not a test-case-writing role. You will build the verification infrastructure behavioral fences, property-based test suites, static analysis pipelines, and agentic QA workflows that allows AI-generated code to ship safely at high velocity. You will work directly with engineering leadership on one of the most consequential transformations in the company's history, with genuine authority to define how quality is engineered here.

Requirements

Strong software engineering fundamentals. You write production-quality code (Python, TypeScript/JavaScript, C#, Java, or Go) and think in systems, not scripts.
Active use of AI coding assistants. Tools like Claude Code, Copilot, or Cursor are part of your daily workflow, and you're genuinely curious about how far they can be pushed you have opinions about where they fail and how to catch it.
An instinct for testing invariants rather than examples experience with property-based tools (Hypothesis, fast-check, QuickCheck-family) is great evidence, but the mindset is what's required.
CI/CD ownership. You have built or owned pipelines, not just consumed them gates, quality signals, and tooling that other engineers rely on.
Static analysis and security tooling familiarity (e.g., Semgrep, SonarQube, CodeQL, Gitleaks) and an understanding of why input/output testing alone cannot catch security or performance rot.
Performance and reliability awareness. Knowledge of performance tuning, capacity planning, and reliability patterns correct output at unacceptable latency or fragility is still a defect.
Strong diagnostic and problem-solving skills for complex systems, and the ability to work collaboratively across development, QA, and operations teams.
A builder's bias. You would rather build a harness that finds a whole class of bugs than manually find one bug.

Nice To Haves

Agentic AI experience. Hands-on with Claude Code or comparable agents generating test scaffolding, accelerating authoring, analyzing coverage gaps, or building agentic workflows of your own.
Experience testing or evaluating LLM-based systems. Evaluation frameworks (evals), model-graded testing, handling non-determinism, and qualitative evaluation (rubrics, rating systems, calibrated human review) where quality can't be reduced to assertions.
Hands-on property-based or generative testing with Hypothesis, fast-check, or QuickCheck-family tools in real projects.
Supply-chain planning, manufacturing, or ERP domain exposure.
Experience characterizing legacy systems (snapshot/approval testing) ahead of large refactors.
Performance and load-testing baselining experience.
SRE or DevOps background on-call experience, chaos/resilience testing, or advocacy for test-driven development and resilience patterns.

Responsibilities

Design and build behavioral fences (characterization harnesses that lock in existing behavior, and acceptance harnesses derived from specifications) that AI-generated code must pass before merge.
Establish system-level behavioral baselines of our core product so agentic refactoring and extension can be verified against observable behavior.
Build and own property-based testing suites (invariant-driven, generative input) targeting the failure modes characteristic of AI-authored code.
Stand up and operate the non-behavioral verification stack: SAST/security scanning, secrets detection, dependency/supply-chain analysis, and performance baselines in CI/CD.
Develop agentic QA workflows AI agents that generate, execute, and triage tests with humans reviewing at the level of strategy and coverage, not individual cases.
Define and report honest coverage boundaries: what is verified, what is sampled, and what is not covered so green dashboards mean what they claim.
Implement observability for quality: logging, tracing, and metrics that surface verification signals, escaped defects, and behavioral drift across the pipeline and in production.
Participate in incident response and post-mortems for quality escapes, feeding every production defect back into the fences, suites, and pipeline as a permanent regression guard.
Automate deployment of the verification infrastructure itself test environments, harnesses, and pipeline tooling.
Develop test strategies for AI-powered product features: prompt regression testing, behavioral boundary testing, output quality evaluation, and regression detection for non-deterministic outputs with observable quality signals defined alongside product and engineering.
Partner with senior product engineers to route AI-generated defect fixes through the verification pipeline with human review at the right altitude.