Description
Televic Rail is seeking one master-level research intern to develop a fleet-level consolidation method for embedded machine-learning models used in accelerometer self-diagnostics on the COSAMIRA Edge platform. The internship combines research, implementation, and experimentation to aggregate learned models across heterogeneous railway trains, detect and reject outliers, ensure robustness to mechanical and environmental variability, validate local learning convergence, and improve fault-detection performance. The work is based at Televic and involves approximately 20% research, 30% implementation, and 50% experimentation.
