Data Science Intern (AI/ML – Reinforcement Learning for Air Traffic Management)

Location: Singapore - Suntec, Singapore

Thales people architect solutions that are relied upon to deliver operational advantage at every decisive moment throughout the mission. Defence and armed forces customers rely on us to deliver the full range of defensive systems for land, sea, and air. From early warning, to threat neutralisation, our platforms cover all levels from very short-range systems, to extended protection across the entire battle-space including Airspace Mobility Solutions, Vehicles and Tactical Systems and Missile Defence, Optronics, and Radar.

Thales established its presence in Singapore in 1973 to support the expansion of aerospace-related activities in the Asia-Pacific region. Throughout the last four decades, the company grew from strength to strength and is today involved in the primary businesses of Aerospace (including Air Traffic Management), Defence & Security, Ground Transportation and Digital Identity & Security. Thales today employs over 2,100 people in Singapore across all its business areas.

About AIR Lab
AIR Lab is a dynamic collaboration between Thales and CAAS, aimed at developing cutting-edge solutions for the aerospace industry, namely in the context of Air Traffic Management (ATM).

We are seeking a hands-on and dedicated Data Engineering Intern to support our Teams. If you are passionate about aviation, data analysis, AI/ML and thrive in a fast-paced, innovative environment, this role is for you.

Position Overview:

AIR Lab is pioneering innovative solutions for ATM and aviation sustainability. As part of our Green Aviation and Artificial Intelligence Team, you will contribute to cutting-edge research and development in AI-automated conflict detection and resolution (CDR). The team’s goal is to create, train, adapt, and improve an AI agent based on reinforcement learning (e.g. Deep Q-Network). As a non-AI module will handle conflict detection leveraging a deterministic approach, the AI model will optimise decision-making based on different types of conflict-trajectory resolution strategies, each offering specific characteristics (hence varying rewards in the AI model’s training) in terms of trajectory efficiency and consequently, its sustainability.

Under the guidance of your Mentor and Team Lead, you will play a pivotal role in enhancing our Data Science and AI capabilities to support the development of innovative aviation solutions.

Responsibilities:

As a Data Science Intern focusing on AI, you will contribute to AIR Lab’s simulation and procedure modelling efforts for trajectory optimisation and conflict resolution in aviation. Your responsibilities include:

1. Understanding the Conflict Detection & Resolution Function
Gain a comprehensive understanding of the existing deterministic conflict detection function and the reinforcement learning-based conflict resolution algorithms. This includes understanding the four types of controller clearances, their associated rewards, and their impact on trajectory optimisation and environmental sustainability.

2. AI Development for Conflict Resolution
Participate in the design, training, and improvement of the AI agent. This will include setting up the Deep Q-Network for reinforcement learning, defining reward mechanisms for each type of clearance, and evaluating the AI's performance in simulation scenarios. Analyse how AI decisions affect flight trajectories and efficiency metrics.

3. Optimising AI Agent Performance
Continuously fine-tune the AI agent by analysing its performance in various scenarios, adjusting hyperparameters, and enhancing its learning process to improve decision accuracy and sustainability outcomes.

4. Scenario Analysis & What-If Modelling
Use the AI agent to run "what-if" scenario analyses on different flight paths and air traffic situations. Help explore various resolutions to conflicts and evaluate their impacts on safety, and trajectory efficiency.

5. Data Collection & Processing
Collaborate with the data engineering team to ensure that relevant aviation data is fed into the AI model. Work on extracting, transforming, and loading (ETL) processes to cleanse and prepare data for simulation and AI training.

6. Collaboration and Communication
Work closely with your Mentor and Team Leads to ensure that your activities are in line with the defined objectives. Share your findings, insights, and recommendations with cross-functional teams.

7. Continuous Learning and Exploration
Stay up to date with the latest advancements in reinforcement learning and AI for aviation. Participate in workshops, training sessions, or other learning opportunities – for instance, those provided by the AIR Lab.

Requirements:

Educational Background: Bachelor's degree in Computer Science, Data Science, Engineering, or a related discipline.
AI/ML Learning Skills: Basic understanding or coursework in reinforcement learning, particularly in Deep Q Networks or similar frameworks.
Programming Skills: Proficiency in Python and familiarity with data science libraries (e.g., TensorFlow, Keras, PyTorch). Experience with data manipulation libraries like Pandas and NumPy is a plus.
AI/ML Knowledge: Understanding of AI/ML models and their application to real-world problems, particularly in optimisation and decision-making. Prior experience in AI/ML projects is beneficial.
Aviation Interest: A strong interest in aviation or air traffic management, with an understanding of how AI can impact sustainability and efficiency in this domain. Prior experience in aviation projects is an advantage.
Problem-Solving Skills: Ability to tackle problems creatively, think critically, and develop innovative solutions.
Collaboration & Communication: Ability to work effectively in a multidisciplinary Team and communicate technical concepts clearly.

Commitment: Able to commit on a full-time basis for ideally 6 months (or a minimum of 5 months) from May/June 2025 onwards.

Learning Outcomes/Benefits:

• Opportunities for professional development and hands-on experience with cutting-edge

• AI and reinforcement learning techniques.

• Collaborative and innovative work environment.

• Chance to work on impactful aerospace projects that contribute to sustainable aviation globally.

At Thales we provide CAREERS and not only jobs. With Thales employing 80,000 employees in 68 countries our mobility policy enables thousands of employees each year to develop their careers at home and abroad, in their existing areas of expertise or by branching out into new fields. Together we believe that embracing flexibility is a smarter way of working. Great journeys start here, apply now!