AI Engineer - Satellite-Based Detection & Prediction Systems

Job Detail

Core AI & Algorithm Development: Design, train, and deploy advanced computer vision models (e.g., CNNs, Vision Transformers, foundational geospatial models) optimized for detecting small, multi-scale targets (ships, boats, debris) within massive, high-resolution satellite imagery. Implement robust probabilistic state-estimation and tracking algorithms (such as Particle Filtering, Particle Smoothing, and Extended Kalman Filters) alongside deep sequential models (LSTMs, GRUs, Transformers) to predict future positions and behaviors of moving targets. Develop multi-modal data fusion strategies to integrate asynchronous, disparate data sources (e.g., combining optical imagery, SAR data, and live transponder telemetry streams) to minimize false positives and overcome environmental limitations like cloud cover. Data Engineering & Geospatial Pipelines: Build scalable, high-throughput ingestion pipelines capable of processing continuous spatial-temporal data streams and converting raw satellite products into cloud-optimized formats (e.g., GeoTIFFs, COGs, Zarr). Architect time-series and geospatial database solutions (e.g., PostgreSQL/PostGIS, TimescaleDB) to store, query, and aggregate billions of historical and real-time coordinates and metadata efficiently. Infrastructure & MLOps Leadership: Lead the deployment, orchestration, and scaling of production AI models within cloud or hybrid-cloud environments using Docker and Kubernetes. Implement end-to-end MLOps frameworks to automate data labeling pipelines, monitor feature/concept drift, handle automated model retraining, and manage model versioning. Write clean, maintainable, and highly optimized code (Python, C++) to handle low-latency processing requirements.