Reinforcement Learning Engineer (RL & Optimization)

20.8k - 28.8k PLN20 800 - 28 800 PLN/ mies./ mies.B2BB2B (netto)

MidZdalnie·WrocławFull-time·B2B

#461522·Dodano 4 dni temu·1

Źródło: justjoin.it

Aplikuj teraz

Tech Stack / Keywords

Reinforcement LearningDeep Reinforcement LearningMathematical OptimizationPython PyTorch Machine LearningsimulationDeep Learning NumPy TensorFlow

Podsumowanie oferty

Develop and improve Reinforcement Learning agents for an Energy Management System serving commercial and industrial facilities. The project combines Reinforcement Learning, mathematical optimization, and energy system simulation to control battery storage and energy flows. The role uses Python and deep learning frameworks such as PyTorch.

Wymagania

Reinforcement Learning:

Commercial, hands-on experience developing, training, and evaluating Reinforcement Learning models.
Practical experience with at least one RL algorithm, such as PPO or DQN.
Ability to independently design RL environments, including reward functions, state representations, action spaces, and constraints.
Experience analyzing RL agent behavior, debugging training issues, and improving policies based on experimental results.

Mathematical optimization:

Solid understanding of mathematical optimization fundamentals, including objective functions and constraints.
Ability to formulate optimization problems involving discrete and continuous decisions.
Understanding of how optimization-based solutions can be used to evaluate and benchmark RL models.

Machine Learning & Software Engineering:

Very good knowledge of Python.
Strong practical experience with PyTorch, TensorFlow, or another deep learning framework.
Very good knowledge of pandas and NumPy.
Practical experience training, testing, evaluating, and comparing Machine Learning models.
Hands-on experience deploying Machine Learning models into production environments.
Experience designing reliable experiments, evaluating model performance, and interpreting results.
Ability to write clean, maintainable, and testable Python code.
Experience with Git.
High level of independence, problem-solving skills, and ownership of technical decisions.
English at B2 level or higher.

Nice to have:

Experience with Reinforcement Learning for control systems, robotics, autonomous systems, or simulation-based decision-making.
Knowledge of energy systems, renewable energy, battery storage, or energy markets.
Experience with Gymnasium or TorchRL.
Experience with linear programming, MILP, or dynamic programming.
Experience developing simulators or simulation environments for RL agents.
Experience with time-series forecasting and time-series data processing.
Knowledge of forecasting architectures such as TFT or TiDE.
Experience with Optuna, DVC, or MLflow.
Experience working with weather, renewable energy production, energy consumption, or electricity price data.
Experience using AI coding assistants such as Claude Code or Codex.
Experience with Docker and/or Kubernetes.

Obowiązki

Design, train, test, and improve Reinforcement Learning agents and models.
Develop mathematical optimization models and use them as benchmarks for RL-based solutions.
Define reward/objective functions, state and action spaces, decision constraints, and RL environments.
Develop and improve simulations of battery storage systems, PV installations, and energy flows.
Evaluate RL policies against optimization-based solutions and baseline control strategies.
Analyze agent behavior, identify weaknesses, and improve model performance through experimentation.
Test and validate models across different installation configurations and energy consumption profiles.
Research and evaluate RL algorithms and optimization approaches to select suitable solutions.
Prepare and process data related to energy consumption, PV production, weather, and electricity prices.
Occasionally support electricity price forecasting models for the day-ahead market.

Benefity

Long-term cooperation.
A technically challenging project combining Reinforcement Learning, mathematical optimization, simulation, and energy systems.
High level of independence and ownership.
Real influence on technical decisions, model architecture, and the direction of the solution.
Opportunity to develop intelligent control systems with a measurable impact on energy costs and renewable energy utilization.

KYOTU Technology

16 aktywnych ofert

Zobacz wszystkie oferty

Aplikuj teraz