New opportunity

Research Scientist - AI for Power Electronics

Hitachi Energy · Raleigh, United States of America

About this role

Hitachi Energy seeks a Research Scientist to develop AI-enabled software tools for designing, optimizing, and validating power electronics systems, including HVDC, STATCOM, and SVC. The role is based on-site at its Research Center in Raleigh, North Carolina.

Responsibilities

Collaborate with global researchers, engineers, and software developers to turn power electronics design challenges into practical tools.
Research design automation using AI, optimization algorithms, and data-driven methods. Develop and validate approaches for design space exploration, performance prediction, and system optimization.
Support component selection, reliability assessment, and scalability improvements; identify technology gaps and help set future research priorities.
Work with universities, national laboratories, and industry partners. Communicate findings through scientific publications, technical reports, and patent applications, while complying with applicable regulations, procedures, and guidelines.

Requirements

Existing authorization to work for Hitachi Energy in the United States.
A Ph.D. in Electrical Engineering focused on power electronics, energy conversion, or a closely related field, with an educational background spanning power electronics and computer science or AI.
Experience in power electronics design, modeling, hardware systems, and simulation or verification, plus hands-on application of AI methods such as supervised or unsupervised learning, optimization, surrogate modeling, or reinforcement learning to engineering problems.
Programming proficiency in Python or a similar language; familiarity with PyTorch or TensorFlow and scientific computing tools; experience with MATLAB, Simulink, PSCAD, and PLECS; and familiarity with hardware design workflows and controller design.
A research record with publications in recognized journals or conferences and effective English communication skills.

The posting mentions learning and career development programs, health and retirement benefits, paid leave, and flexible work arrangements; the stated work arrangement for this role is on-site.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

Power electronicsArtificial intelligenceDesign automationOptimization algorithmsDesign space explorationPerformance predictionReliability assessmentSupervised learningUnsupervised learningSurrogate modelingReinforcement learningPythonPyTorchTensorFlowScientific computingMATLABSimulinkPSCADPLECSHardware designController designSimulationResearch communication

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

AI-enabled design tools
The role develops software tools for designing, optimizing, and validating power electronics systems such as HVDC, STATCOM, and SVC. Research in design automation must translate into practical design workflows.
Validated engineering methods
Develop and validate approaches for design space exploration, performance prediction, and system optimization, drawing on power electronics modeling, hardware knowledge, and simulation or verification.
System-level design decisions
Support component selection, reliability assessment, and scalability improvements, and identify technology gaps that can inform future research priorities.
Cross-functional delivery
Work with global researchers, engineers, and software developers to turn engineering challenges into practical, user-focused tools; collaborate with universities, national laboratories, and industry partners.
Research communication and compliance
Share findings through publications, technical reports, and patent applications while following applicable internal and external regulations, procedures, and guidelines.

Likely challenges to think through

Balancing model speed and engineering confidence
Preparation hypothesis: Faster AI-based exploration or prediction may be useful only if results remain credible for power electronics design. Prepare an example explaining validation against simulation, verification, or hardware evidence and how you handled error or uncertainty.
Turning research prototypes into usable tools
Preparation hypothesis: A promising algorithm may not fit an engineer’s workflow or scale across applications. Prepare to discuss how you would define a design problem with users, choose practical performance measures, and iterate on a prototype.
Reconciling competing design goals
Preparation hypothesis: Optimization, component selection, reliability, and scalability may pull decisions in different directions. Prepare a concrete way to frame constraints and trade-offs rather than presenting a single metric as sufficient.

Practice your answers

Technical practice

How would you model an HVDC, STATCOM, or SVC system for a design study?

What to coverSuggested preparation: Explain which electrical and control behaviors your model must capture for the question at hand. Discuss the trade-off between model fidelity, simulation cost, and usefulness for design decisions.

How would you verify a power electronics model before using its results to guide a design?

What to coverSuggested preparation: Describe checks against known operating conditions, simulations, or available hardware evidence. Explain how you would investigate discrepancies and communicate the model’s limits.

How would you turn a power electronics design workflow into an AI-enabled software tool?

What to coverSuggested preparation: Start with a specific engineering decision, its inputs and outputs, and the data needed to support it. Explain how you would prototype, validate, and make the tool useful to engineers.

How would you use a surrogate model for performance prediction during design space exploration?

What to coverSuggested preparation: Discuss training-data coverage, error measurement, and testing on designs not used for fitting. Explain when the surrogate would need confirmation with a higher-fidelity model.

How would you formulate an optimization problem for a power electronics system?

What to coverSuggested preparation: Identify candidate design variables, objectives, and engineering constraints without assuming a particular project specification. Explain how you would compare solutions when performance goals conflict.

When would you choose supervised learning, unsupervised learning, or reinforcement learning for an engineering problem?

What to coverSuggested preparation: Compare the data, feedback, and validation each approach requires. Explain why a simpler optimization or modeling method might be preferable when it addresses the problem reliably.

How could a design tool support component selection and reliability assessment?

What to coverSuggested preparation: Explain how component characteristics and modeled operating conditions could inform a selection. Discuss uncertainty in reliability estimates and how you would avoid presenting predictions as guarantees.

What would you test before scaling a research prototype into a reusable design tool?

What to coverSuggested preparation: Consider computational cost, repeatability, and behavior across different design cases. Explain how you would detect where the approach no longer performs adequately.

How would you connect controller design and hardware workflows with simulation-based research?

What to coverSuggested preparation: Describe the role of controller behavior in evaluating a power electronics design and how simulation results could inform hardware decisions. If relevant to your experience, explain how you would use MATLAB, Simulink, PSCAD, or PLECS, distinguishing each tool’s purpose.

How would you identify a research gap in AI-assisted power electronics design?

What to coverSuggested preparation: Describe how you would establish a baseline, define a measurable limitation, and design an experiment to test an improvement. Address what evidence would justify further research rather than immediate tool development.

Behavioural practice

Tell me about a time you translated an engineering challenge into a usable tool.

What to coverSuggested preparation: Use a real situation, your actions to understand users’ needs and build the solution, and the result. Clarify your contribution and how you assessed whether the tool helped.

Describe a time you worked across research, engineering, and software disciplines.

What to coverSuggested preparation: Give the situation, how you resolved differences in terminology or priorities, and the result. Focus on actions that made the collaboration effective.

Tell me about a research collaboration with an external partner.

What to coverSuggested preparation: Use an actual example involving an academic, laboratory, or industry partner if you have one. Explain the shared objective, your actions, and the outcome without implying the role requires a particular partnership model.

Describe a time you communicated a technical finding to different audiences.

What to coverSuggested preparation: Explain the situation, how you adapted a publication, report, or presentation to its audience, and the result. Include how you conveyed limitations or uncertainty.

Tell me about a time evidence changed your research priorities.

What to coverSuggested preparation: Describe the initial direction, the evidence that exposed a gap or limitation, and what you did next. State the outcome and how you communicated the change to collaborators.

Sharpen your knowledge

Topics to prepare

Grid-connected power electronics
The posting names HVDC, STATCOM, and SVC as systems the tools will support. For revision, sketch each system’s purpose, major conversion stages, and design variables that could affect grid performance.
AI methods for engineering problems
The posting calls for hands-on application of supervised or unsupervised learning, optimization, surrogate modeling, or reinforcement learning. For revision, compare when two of these approaches would suit performance prediction versus design exploration, including their data and validation needs.
Design automation and optimization
The role focuses on design space exploration, performance prediction, and system optimization. As a preparation exercise, formulate a power-electronics design problem with decision variables, constraints, competing objectives, and a way to check the proposed optimum.
Reliability and scalable design
The role supports component selection, reliability assessment, and scalability improvements. For revision, work through how changing a component choice could affect performance, reliability, and the usefulness of a design tool at larger scale.
Model-to-hardware validation
The posting requires experience with modeling, hardware systems, simulation or verification, and familiarity with controller design. For revision, outline what evidence would distinguish a promising simulated controller result from a design ready for hardware evaluation.

Skills to revise

Build a small AI prediction prototype
The posting asks for Python or similar programming, familiarity with PyTorch or TensorFlow, and scientific computing tools. As practice, use one named framework to fit a small surrogate model to simulated power-electronics results, then report prediction error on held-out cases.
Run and interrogate a simulation
The posting names MATLAB, Simulink, PSCAD, and PLECS. For practice, use a named tool you can access to vary a converter or controller parameter, record performance outputs, and explain how you would verify the result; this is a suggested exercise, not a specified employer test.
Prototype a design-space search
The role develops and validates approaches for design exploration and optimization. As practice, write a small search routine over simulated design choices, enforce explicit constraints, and compare its selected design with a baseline.
Present research findings clearly
The role communicates findings through publications, technical reports, and patent applications and works with engineering and external research partners. For practice, prepare a brief technical summary of a relevant study or prototype covering the question, method, validation, limitations, and practical design implications.

Your preparation checklist

Map the stated requirements to evidence
Prepare a one-page reference to your Ph.D. focus, power electronics and AI education, relevant experience, publications, and existing U.S. work authorization. These are stated requirements; use only accurate evidence.
Build a power-electronics project walkthrough
Choose a real project and explain the system, design and hardware constraints, modeling approach, simulation or verification, controller considerations, and measured or simulated results.
Explain an applied AI method end to end
Prepare a concrete engineering example covering the data, method, baseline, validation, limitations, and outcome. The posting names optimization, learning, and surrogate modeling as relevant methods; discuss only those you have used.
Practice a design-tool exercise
As interview preparation, sketch how you would turn an HVDC, STATCOM, or SVC design challenge into a tool for design-space exploration or performance prediction. Include inputs, outputs, validation checks, and how engineers would use it; this is a practice exercise, not a stated interview task.
Prepare a technical-tool inventory
List specific work you can discuss in Python or a similar language, scientific computing, ML frameworks, MATLAB, Simulink, PSCAD, PLECS, and hardware design workflows. Mark gaps honestly rather than implying experience with every named tool.
Select research communication examples
Choose a publication or technical report and prepare a short account of your contribution, evidence, limitations, and how you communicated results to collaborators. The role also mentions patent applications and external research partners.
Write questions about research priorities and validation
Bring questions about current gaps in design tools, how prototypes are validated against engineering needs, collaboration across research and software teams, and expectations for work at the on-site Raleigh Research Center.

Availability last checked: Nov 4, 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in United States of America. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…