Performance and quality testing of Foundation Models On Device (FMOD) is mission-critical to Argmax. We set the expectations of our Enterprise customers with SLAs that are backed by end-to-end testing on real devices at scale. Due to the growing number of projects and customers, we are now hiring an ML Software QA Engineer to consolidate all of our testing work under a unified and scalable testing infrastructure and lead the testing function moving forward.
Responsibilities
- Unify the testing strategy and infrastructure across all customer-facing projects
- Manage and scale our in-house test hardware lab
- Evaluate the integration of external on-device testing services while scaling our in-house lab
- Automate regression testing for new operating systems and new hardware releases
- When necessary, conduct manual testing for customer-critical workloads
- Root cause test failures to data integrity issues or software bugs
- Convert bug reports into test cases and identify proactive mitigation strategies for related issues
Qualifications
- 3 years of relevant industry experience
- Fluency with at least 2 languages among Swift, Python, C/C++
- Familiarity with modern ML software frameworks (e.g. PyTorch, Core ML, MLX)
- Familiarity with iOS, macOS, Linux, and Android testing frameworks and environments
Preferred Qualifications
- Proven track record of leading QA engineers
- Proven track record of setting up hardware labs for on-device testing
Perks
- Top-of-market equity at a fast-growing early-stage startup with a unique mission
- 3 days per week in office from Palo Alto, CA or Manhattan, NY
- Palo Alto office offers comprehensive on-site amenities
- Remote possible for industry leader exceptional candidates
- Platinum-tier healthcare with 90% employer contribution
- 401(k) match
- Quarterly team building offsite
About Argmax:
Our mission at Argmax is to empower developers and enterprises everywhere, who are eager to migrate inference from cloud servers to user devices, or to turbocharge their existing on-device inference. We believe that achieving this at scale will pull the future forward when AI is private and convenient for users, and profitable and maintainable for companies building it.