Data Scientist

About the role

Thales is seeking a Data Scientist to drive AI‑powered solutions for air traffic management in Singapore. You will work on cutting‑edge machine‑learning models, reinforcement‑learning agents and large‑language‑model pipelines to optimise aeronautical operations.

Key responsibilities

• Design and conduct exploratory data analysis to uncover patterns and optimisation opportunities in air traffic data.
• Develop, train and deploy ML models (classification, regression, sequence prediction) using PyTorch, TensorFlow or Scikit‑learn, including transformer‑based spatial‑temporal architectures.
• Build reinforcement‑learning agents (DQN, PPO, Actor‑Critic) and apply them in simulated or real‑world environments via OpenAI Gym or custom setups.
• Create and optimise Retrieval‑Augmented Generation (RAG) pipelines grounded on domain‑specific documentation.
• Evaluate large‑language‑model outputs for hallucination and groundedness, establishing domain‑specific benchmarks.
• Automate end‑to‑end ML pipelines with Kubeflow, Airflow or similar orchestration tools.
• Design reproducible workflows for data preprocessing, model training, evaluation and deployment.
• Integrate models into scalable APIs and deploy them in cloud‑native environments using Docker and Kubernetes.
• Monitor model performance, manage drift, and maintain experiment tracking with MLflow or Weights & Biases.
• Collaborate with DevOps and backend engineers to ensure seamless system integration.

Required profile

• Bachelor’s degree in Computer Science, Information Technology or a related field (Master’s preferred).
• Strong analytical mindset with experience in domains that have limited historical data.
• Aeronautical or air‑traffic‑management knowledge is a strong plus.

Required skills

• Python programming.
• PyTorch, TensorFlow, Scikit‑learn.
• Transformer and spatial‑temporal modeling.
• Reinforcement‑learning algorithms (DQN, PPO, Actor‑Critic) and OpenAI Gym.
• RAG pipeline development.
• ML orchestration tools such as Kubeflow or Apache Airflow.
• Containerisation and orchestration with Docker and Kubernetes.
• Experiment tracking with MLflow or Weights & Biases.