Caroline Silva

Senior ML Engineer & Applied Researcher with 8+ years of experience industrializing AI/ML systems — from computer vision research to ML pipelines and platform tooling. I combine the rigor of a PhD in Computer Vision with hands-on AI systems engineering, bridging research experimentation and production deployment through ML workflows, model monitoring, and MLOps practices. My work focuses on real-world constraints such as noisy data, limited datasets, latency budgets, and long-term model reliability. Industrial AI Projects • Edge AI for Maritime Surveillance Designed edge-deployable ship detection models for satellite imagery using NAS, optimizing recall, latency, and FLOPs trade-offs. Implemented production drift monitoring for covariate and concept drift. • Industrial Operator Safety Monitoring Evaluated and deployed segmentation models including DFormerV2, SAM2, and MobileSAM for real-time operator detection in industrial risk zones. • Aerospace Electronics Defect Detection Developed a hybrid inspection pipeline combining U-Net segmentation with Procrustes-based geometric analysis for robust aerospace defect detection. AI Platform & Infrastructure At ActiveEon, I developed the AI/ML layer of the ProActive Orchestration platform, contributing 100+ customizable workflows for ML, AutoML, Drift Detection, and Model-as-a-Service applications. Also contributed to the Python SDK through the @task and @job decorator system for scalable ML workflow orchestration. Research & Publications Author of 4 peer-reviewed publications and speaker at ICPR, VISAPP, OW2con, SophiaConf, and WAICF. Research Interests Computer Vision · Applied AI Systems · Neural Architecture Search · Drift Detection · Deep Learning Tech Stack Python · PyTorch · ONNX · Triton · FastAPI · MLflow · Docker · AWS · Prometheus/Grafana Particularly interested in roles that bridge resear