Intern - AI Systems

• Assist in defining the overall AI architecture for aerostat stabilisation, including sensor fusion, fault detection, and predictive control loops.
• Evaluate and implement ML frameworks such as TensorFlow, PyTorch, and ROS (Robot Operating System).
• Support in setting up the data acquisition and preprocessing pipeline from IMU, GNSS, wind sensors, and tether sensors.
• Contribute to the development of Reinforcement Learning (RL) or Model Predictive Control (MPC) algorithms for:
o Dynamic attitude stabilisation (pitch, roll, yaw).
o Real-time tether tension and azimuth compensation.
o Active damping under wind disturbances.
• Implement simulation models in MATLAB/Simulink or Python to evaluate control algorithms before hardware integration.
• Work with electrical/mechanical teams to integrate control algorithms with motor drive and winch systems.
• Participate in Hardware-in-the-Loop (HiL) testing to validate AI control performance.
• Collect and label flight-test data for continual model improvement and anomaly analysis.
Anomaly Detection & Predictive Insights
• Develop lightweight AI models for detecting abnormal tether loads, motor temperature, or drift instability.
• Implement predictive analytics dashboards for real-time health monitoring and performance alerts.
Documentation & Reporting
• Prepare detailed technical documentation on algorithms, datasets, and model evaluation metrics.