Principal Applied Scientist

Siemens•Washington, DC

12h•$167,178 - $226,236•Hybrid

About The Position

Siemens builds the systems that power the physical world, including factories, power grids, buildings, trains, and hospitals. Industrial and physical AI presents a significant opportunity in applied AI, and it is a challenging area to master. There is a generation of AI-powered products to be built. We are an applied science organization focused on developing the underlying science for these products. Our work lies at the intersection of machine learning research, real-world data, and production systems operating in industrial environments. As a Principal Applied Scientist, you will lead the scientific direction for a major capability area within the pod. You will tackle ambiguous problems, guiding them through hypothesis formulation, method selection, experimentation, and ultimately into models deployed in production. You will be hands-on with the modeling process and accountable for its rigor and outcomes. This is a senior individual contributor role. You will collaborate closely with engineers and product managers who are responsible for translating your work into products.

Requirements

8+ years in applied machine learning, AI research, or data science, with models that shipped to production and made an impact
Strong foundation in machine learning theory and practice across training, evaluation, and deployment
Demonstrated experience taking research from a paper or prototype into a model that runs reliably in production
Proficiency in Python and modern ML frameworks and toolchains
Strong partnership track record with engineering teams on data, training infrastructure, and inference
Clear written and verbal communication with engineers, product managers, and senior leaders

Nice To Haves

Experience applying ML in industrial or physical domains: manufacturing, automation, robotics, energy, mobility, infrastructure, healthcare
Deep expertise in one of: multimodal ML, generative AI, retrieval augmented generation, agentic workflows, time series, control, or planning
Scientific ML for physical systems: surrogate modeling, operator learning, physics-informed ML, geometry-aware ML, differentiable simulation, AI for semiconductor/EDA
Hands-on experience building evaluation pipelines, running online experiments, or instrumenting production monitoring for a model you owned
Publications, patents, open source contributions, or significant internal technology transfers
Experience mentoring more junior scientists and engineers
Experience working with globally distributed research, product, or engineering organizations

Responsibilities

Own one or more scientific capability areas end to end, for example perception, computer vision, language and agents, time series, control, planning, or evaluation
Take problems from ambiguous product or system requirements through clear research questions, hypotheses, and success metrics
Lead applied research projects: literature review, method selection, experimentation, ablation, error analysis, and productization
Build and run the evaluation pipelines for the work you own: offline metrics, online experiments, robustness testing in industrial conditions
Work with engineers to take models into production grade pipelines: data readiness, training infrastructure, inference, observability
Make scientific tradeoffs in front of engineers and product managers, with evidence, and translate them into decisions the team can act on
Identify and de-risk scaling challenges in your area: data quality, model drift, latency, throughput, cost, safety
Raise the bar on experimentation rigor, reproducibility, and documentation across the team
Apply responsible AI practices in your work: bias detection, model risk management, human in the loop controls