Applied AI Scientist, Technology Partnerships (Senior/Principal)

ABOUT SYGALDRY Sygaldry Technologies is building quantum-accelerated AI servers to exponentially speed up training and inference for AI. By integrating quantum and AI, we're accelerating the path to superintelligence, and addressing the problem of rising compute costs and energy bottlenecks. Sygaldry AI servers combine multiple qubit types within a single, fault-tolerant architecture to deliver the combination of cost, scale, and speed necessary for advanced AI applications. We pioneer new domains in physics, engineering, and AI, tackling the hardest challenges with a grounded, optimistic, and rigorous culture. We're looking for individuals ready to define the intersection of quantum and AI and drive its profound global impact. THE ROLE Our AI & algorithms team develops quantum approaches to training, inference, and reasoning, including quantum-native generative modeling. Much of this work runs today on classical hardware at small scale, built so that results transfer directly to quantum processors as they come online. Increasingly, that research meets real problems. Partner organizations bring workloads from their own domains, and someone has to carry our methods into each one and find out how well they hold. That is this role. You are a computational scientist who takes scoped collaboration and does the technical work: understand the partner's problem well enough to know where our methods can genuinely help, prototype, benchmark against what they already trust, and be the technical voice in the room when results are discussed. Our partnerships managers own the relationships and run the programs; our research scientists develop the algorithms. You produce the evidence. WHAT YOU'LL WORK ON Applied Work in Partner Domains - Assess where our quantum-native generative methods apply to a specific domain: which algorithm, under what assumptions, in which regime it wins, and what would count as evidence - Build and run the proof-of-concept work, and adapt research-team implementations to relevant domains - Validate against the reference methods the partner already trusts, whether that's molecular dynamics, Monte Carlo, a classical solver, or their production model - Present and refine results in technical working sessions, and carry what you learn -- model scales, data characteristics, evaluation criteria -- back into decisions about what we develop next Benchmarking & Evidence - Produce evidence by whichever route fits the question: approximate simulation where a structured representation scales, circuit emulation where exact small-scale results matter, and analytic resource models to reach past what either can run - Design benchmarks that hold up: quantum, classical, and hybrid compared under realistic assumptions, against the strongest classical method rather than a convenient one - Use our internal modeling and simulation environment to show a partner what their own workload would look like on our architecture - Extend those models to new algorithm workloads, and produce the charts, comparisons, and reproduction artifacts behind our published and partner-facing results YOU MAY BE A GOOD FIT IF YOU: - Are a scientist who ships: you write code daily, and your results are reproducible by someone else - Are quantitatively deep enough to judge whether a result is right, not only whether the pipeline ran - Can hold a technical conversation with a domain expert in a field you didn't train in, and come out knowing where the real bottleneck is - Value rigor: you are comfortable assessing where quantitative methods help, where they don't, and what evidence distinguishes the two - Move fast, and can carry several engagements in parallel without letting any of them go stale STRONG CANDIDATES MAY HAVE - Advanced training (MS, PhD) in a computational field -- physics, chemistry, applied mathematics, computational biology, engineering -- or equivalent depth built in industry - Strong Python and scientific computing: linear algebra, ODE/SDE so…