Looking for a job

Marco Uccelli

@marcouccelli

Message

I build AI systems for computer vision, Earth Observation, and space programmes.

Italy

Message

What I'm looking for

I'm open to full-remote opportunities where I can build and deploy AI systems across computer vision, Earth Observation, LLM applications, and institutional space programmes.

I'm building AI solutions at Starion Group for institutional space programmes including ASI's VIPER and ESA's REX. My work spans end-to-end detection and super-resolution for Earth Observation and space imagery, plus LLM pipelines for spacecraft anomaly-report annotation.

Previously, I developed predictive-maintenance, anomaly-detection, quality-prediction, and parameter-calibration systems for industrial machinery at Links Management and Technology and DSTech. I also delivered explainable AI monitoring dashboards using SHAP and PowerBI.

At TopNetwork, I trained and deployed deep-learning models on NVIDIA Jetson hardware for warehouse access control, parcel monitoring, and real-time video surveillance. I work hands-on across data pipelines, model deployment, RAG, agentic workflows, and LLM fine-tuning.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

Starion Group S.p.A.

Aug 2023 - Present (3 years)

Developing detection and super-resolution models end-to-end for a model-agnostic processing chain handling Earth Observation and space imagery. Built an LLM pipeline for automated annotation of spacecraft anomaly reports using constrained decoding and few-shot retrieval on a locally-hosted 32B model.

Data Scientist

Links Management and Technology S.p.A.

May 2023 - Aug 2023 (3 months)

Developed an anomaly-detection model estimating Remaining Useful Life (RUL) of industrial machines for predictive maintenance.

Anomaly Detection Predictive Maintenance Python scikit learn PyTorch TensorFlow Keras Pandas NumPy

Data Scientist

DSTech S.r.l.

Jan 2022 - Apr 2023 (1 year 3 months)

Developed predictive-maintenance models with explainable-AI techniques (SHAP) and delivered a monitoring dashboard (PowerBI) for model behavior and anomaly tracking. Built a real-time quality-prediction and parameter-calibration system for industrial machinery, optimizing output quality while reducing energy consumption.

Predictive Maintenance Explainable AI SHAP Power BI Python scikit learn PyTorch TensorFlow Keras Pandas NumPy

AI & Machine Learning Developer

TopNetwork S.p.A.

Jul 2020 - Jan 2022 (1 year 6 months)

Trained and deployed deep-learning models on edge hardware (NVIDIA Jetson), optimizing inference latency and efficiency with TensorRT and TFLite. Built face-recognition access control and BLE-sensor parcel monitoring for warehouse management, and an object-detection pipeline for real-time video surveillance.