Director, Software Quality Engineering

At d-Matrix, we are focused on unleashing the potential of generative AI to power the transformation of technology. We are at the forefront of software and hardware innovation, pushing the boundaries of what is possible. Our culture is one of respect and collaboration.

We value humility and believe in direct communication. Our team is inclusive, and our differing perspectives allow for better solutions. We are seeking individuals passionate about tackling challenges and are driven by execution.  Ready to come find your playground? Together, we can help shape the endless possibilities of AI. 

Software quality engineering answers whether the d-Matrix's stack, firmware, compiler, runtime, and inference serving actually work, automatically, continuously, and with evidence. We're hiring a hands-on director or senior director to lead a team of eight through a strategy reset, with an automation-first mandate where quality is measured, not assumed.

Santa Clara, CA (Hybrid 3 days a week in office)

Quality at d-Matrix is not a gate at the end of the pipeline. A model running on Corsair passes through the firmware, the compiler, the runtime, and the serving stack before a single token comes back. A regression anywhere in that chain shows up as a wrong answer, a slow answer, or an answer that was right yesterday. The Software Quality Engineering team answers the question of whether the whole stack actually works and answers it automatically, continuously, and with evidence.

You will lead the Software Quality Engineering team, which covers that entire range: firmware, compiler and runtime, and inference serving. Eight engineers are already in place, along with real test content and real institutional knowledge about how this stack breaks. The team has been without a permanent leader and needs direction reset rather than rebuilding from zero. Your first job is an honest assessment of what exists, what to keep, what to retire, and where the coverage gaps actually are, followed by a strategy the rest of engineering believes in.

Automation-first is the mandate, not an aspiration. Manual effort belongs in exploratory testing and new-feature bring-up. Anything that needs to run more than a handful of times gets automated or gets deleted.

What makes quality hard here?

This is the part of the job worth taking seriously and the reason it is more interesting than most quality leadership roles:

• Correctness is numerical, not boolean. Quantized inference does not produce bit-identical results. Someone has to define what "close enough" means against reference implementations, build tolerance-aware comparison into the test system, and defend that definition when a model's output drifts.

• Performance is a correctness criterion. We sell latency and efficiency. A double-digit latency regression is a bug of the same severity as a wrong answer, and the test system has to catch it with the same rigor.

• The target moves underneath you. New silicon revisions, firmware updates, and compiler changes land continuously. Test infrastructure that assumes a stable platform will not survive contact.

• Test capacity is finite and physical. Accelerators are scarce and expensive. A real strategy needs tiers, unit, simulated, single-card, multi-card, and full rack, with deliberate choices about what runs where and how often.

• Failure attribution crosses layers. A red test could be a compiler bug, a firmware regression, a bad board, or a flaky test. Making that distinction fast and automatically is most of the value this team creates.

This role is not:

• Building a large manual test organization or managing an outsourced test vendor as the core strategy

• Post-silicon hardware validation. Firmware software validation is in scope; silicon bring-up is not

• Writing unit tests on behalf of development teams. Developers own their own unit tests; this team sets the standard and provides the tooling

• Quality strategy spanning firmware, compiler and runtime, and the inference serving stack, and the test tiering that implements it

• Test frameworks and harnesses: the tooling engineers write tests against, designed for reuse across layers

• Automated functional, integration, and end-to-end suites, from firmware and compiler output through full serving deployments

• Accuracy and numerical regression detection, including reference comparison and tolerance policy

• Performance regression detection and benchmarking hygiene

• Release readiness criteria and quality gates: what blocks a release, stated in measurable terms

• Defect triage flow, escape analysis, and the quality metrics the engineering organization runs on

• Leading, growing, and hiring into a team of eight, and setting the standard for how the rest of the engineering tests its own work

• Experience leading a software quality or test engineering team, while staying hands-on enough to write framework code and review tests yourself

• A track record of taking over an existing team and changing its direction, including growing engineers into new skills rather than replacing them

• Strong programming ability in Python and/or C++. You have built test frameworks, not just written test scripts against someone else's

• Systems software testing depth: firmware, compilers, runtimes, drivers, distributed systems, or comparable. Experience testing things without a user interface

• Hardware and software integration testing, including the failure modes that come with testing against real devices

• Designing test strategy under real resource constraints, where you cannot simply run everything on every commit

• Integrating test suites into CI and defining promotion and release gates that teams actually respect

• A data-driven approach to coverage and risk, and the judgment to know when a quality metric is measuring the wrong thing

• Credibility with developers. This function succeeds by making engineers faster and more confident, not by blocking them

• ML or inference domain knowledge: quantization, numerics, model accuracy evaluation

• Familiarity with inference serving stacks and the benchmarks used to evaluate them

• Experience with Kubernetes-based test environments

• Hardware-in-the-loop testing, device farms, or lab automation

• Practical, skeptical experience applying AI to test generation, triage, or failure clustering - we are interested in what actually works

Equal Opportunity Employment Policy

d-Matrix is proud to be an equal opportunity workplace and affirmative action employer. We’re committed to fostering an inclusive environment where everyone feels welcomed and empowered to do their best work. We hire the best talent for our teams, regardless of race, religion, color, age, disability, sex, gender identity, sexual orientation, ancestry, genetic information, marital status, national origin, political affiliation, or veteran status. Our focus is on hiring teammates with humble expertise, kindness, dedication and a willingness to embrace challenges and learn together every day.

d-Matrix does not accept resumes or candidate submissions from external agencies. We appreciate the interest and effort of recruitment firms, but we kindly request that individual interested in opportunities with d-Matrix apply directly through our official channels. This approach allows us to streamline our hiring processes and maintain a consistent and fair evaluation of al applicants. Thank you for your understanding and cooperation.