About the role
SOTI is seeking a Data Scientist, Aerospace to lead end-to-end AI/ML work on autonomous drones and robotics projects. You will own data strategy, modelling, integration, evaluation, deployment, and monitoring across multi-sensor data pipelines.
The role emphasizes practical implementation, collaboration with cross-functional teams, and mentoring junior contributors in a fast-paced, innovative environment.
Lead defined AI/ML work packages spanning data strategy, modelling, integration, evaluation, deployment, monitoring, and technical documentation. Design perception and sensor-fusion pipelines combining RGB, thermal, depth, LiDAR, radar, IMU, GNSS, telemetry, and environmental data. Select and implement appropriate methods for object detection, segmentation, tracking, pose estimation, localization, mapping support, anomaly detection, forecasting, and decision support. Define dataset coverage plans and simulation or synthetic-data strategies for rare events, difficult environments, and safety-related scenarios. Lead error analysis, ablation studies, uncertainty evaluation, robustness testing, and model-performance trade studies. Optimize models for GPU, NPU, CPU, or embedded platforms using quantization, pruning, compilation, batching, and architecture selection as appropriate. Coordinate integration with autonomy, controls, embedded, cloud, robotics middleware, ground-control, and user-application teams. Establish reproducible training, versioning, experiment-tracking, validation, and release practices for assigned solutions. Investigate data drift, sensor failures, model degradation, false detections, localization errors, and field-test observations using structured root-cause analysis. Mentor Associate, interns, and research contributors. 4–6 years of professional experience in data science, AI/ML, perception, sensor fusion, or related engineering. Demonstrated ownership of models or data products from requirements through integration and validation. Strong experience with Python, deep-learning frameworks, data pipelines, experiment tracking, version control, and Linux environments. Practical experience with multi-modal data, computer vision, sensor fusion, robotics logs, edge inference, or simulation with ROS2 framework. Understanding of Kalman filtering, Bayesian estimation, tracking, coordinate transformations, time synchronization, Path planning, navigation and uncertainty is strongly preferred. Experience with MLOps, containers, CI/CD, model registries, cloud platforms, or embedded deployment is an asset.