Senior Research Engineer, Data Science

The Senior Research Engineer will conduct research and development for the electricity market. The work covers forecasting models for power generation and demand at power plants and demand sites, price prediction models for JEPX spot, imbalance, and balancing markets, and optimization algorithms for trading plans that combine demand sites and storage facilities. The role spans investigation, design, implementation, testing, production operation, revenue-effect analysis, and continuous improvement, with AI Agents used throughout the development and operations workflow.

The engineer is expected to turn research ideas into reproducible, production-quality code and operate the resulting algorithms in services. Responsibilities include researching and developing generation and demand forecasting models, researching and developing electricity-market price forecasting models, implementing algorithms at product quality, analyzing the revenue effects of algorithms in operation, and continuously improving the development and operations process through the use of AI Agents.

Required qualifications include daily use of AI Agents such as Claude Code or Codex for research, design, implementation, testing, and review, together with the ability to independently verify important assumptions, calculations, and deliverables. Candidates must have experience decomposing ambiguous business issues into research questions and designing evaluation criteria and validation plans. They must also have research or practical experience as the primary developer of models or algorithms in at least one of time-series forecasting, probabilistic forecasting including quantile forecasting, or mathematical optimization. The role requires the ability to investigate papers and prior research and translate them into reproducible algorithm implementations and validated results, experience implementing backend services with Python, experience analyzing data with analytical databases such as BigQuery or Snowflake, and experience developing code with GitHub. Japanese communication with business members must be possible without difficulty.

Preferred experience includes taking time-series or probabilistic forecasting models through production operation, connecting improvements to forecasting or optimization models with business results such as revenue or cost and validating their effects, and experience with the electricity or energy market. The expected working approach is to commit to the full improvement cycle from specification planning through production operation, reason from business objectives about what should be done, and act effectively in a fast-changing environment.

The technology stack is Python, TypeScript, Terraform, Google Cloud including BigQuery, Dataform, and Cloud Run, and PostgreSQL. The company is a startup founded in November 2024 and has a team of approximately a dozen people.

The work location is in Shimbashi, Minato, Tokyo, and a remote-work system is available.