Description : Responsibilities :
- Design predictive models for trajectory forecasting, traffic participants behavior, and crossing probabilities.
- Develop risk scoring mechanisms using time-shifted risk prediction and sliding time windows.
- Implement multi-agent reinforcement learning (MARL) frameworks to simulate and train cooperative behaviors.
- Work with simulation teams to integrate ground truth scenarios and replayable datasets.
- Build scoring algorithms for different data dimensions based on the severity and impact.
- Evaluate model performance using precision, recall, and event-level accuracy.
- Collaborate with data engineers to define feature pipelines and streaming inputs.
Requirements :
3+ years of experience in applied data science, preferably in real-time or simulationbased environments.Strong proficiency in Python, NumPy, Pandas, and deep learning frameworks like PyTorch or TensorFlow.Experience with time-series analysis, Bayesian models, or probabilistic forecasting.Understanding of reinforcement learning, especially multi-agent settings.Knowledge of vehicle kinematics, trajectory forecasting, or intelligent transportation systems.Nice To Have :
Experience with simulation environments like CARLA, SUMO or VISSIM simulation data.Prior work on ADAS, or smart city risk management.Familiarity with CEP engines or event stream analytics tools.Understanding of data fusion from camera, LiDAR and other infrastructure inputs.Must Have :
Python- high level coding exp is a must.Traditional MLAWS, Databricks, airflow, GitlabLLM, Gen(ref : hirist.tech)