Research Engineer Intern

This part-time Research Engineer internship focuses on algorithm research and development for the electricity market. The role covers solar-power generation forecasting, electricity-market price forecasting for JEPX spot and imbalance markets, and optimization of trading and bidding plans that incorporate batteries.

Day-to-day work includes improving and validating solar-generation forecasting models, including time-series and quantile forecasting; researching and implementing models for electricity-market prices and imbalance; researching and implementing optimization algorithms for battery operation and bidding plans; backtesting with historical data; evaluating models using accuracy and revenue metrics; and reviewing papers and prior research before translating the findings into implementations. Tasks are prepared with defined problem settings, data, and evaluation metrics. Work begins with separated, well-defined tasks, with the scope expanding according to progress and suitability.

The named technology stack includes Python, pandas, polars, scikit-learn when needed, PyTorch, BigQuery, and GCS. The development environment is designed for use of AI Agent tools, including Claude Code and Codex.

Required qualifications are experience implementing Python algorithms for machine learning or mathematical optimization. This experience may come from research, personal development, competitions, or other settings. Candidates must also have experience developing code with AI Agent tools and be able to read papers or prior research and implement the relevant ideas independently. Japanese communication must be possible without difficulty.

Preferred experience includes time-series forecasting, probabilistic or quantile forecasting, mathematical optimization implementation, achievements in Kaggle or competitive programming, and data analysis using BigQuery or SQL. Interest in the electricity or energy market is also welcomed. The role suits people who form and investigate their own hypotheses, use AI Agents as development tools, and can work independently in a remote and asynchronous setting.

The work uses real electricity-market data and evaluates results by accuracy and revenue. Direct mentoring is available from an engineer responsible for the process from forecasting through bidding and actual operations. There is a path to employment as a regular employee for new graduates or experienced hires.

The workplace is in Minato-ku, Tokyo, and remote work is available. The position is an hourly part-time job with pay of JPY 1,500 to 2,000 per hour. The expected workload is approximately 10 or more hours per week, with shifts between 9:00 and 18:00; working hours and days are negotiable. The probationary period is within three months, and applicants able to work for at least three months are particularly welcome. Commuting expenses are provided, and no insurance enrollment is provided.