Eduarda Pereira

An adaptive urban traffic management system built on top of SNMP/INMF. Features micro-simulation, real-time control, custom Fernet cryptographic hardening, and live visualization dashboards.

An adaptive predictive maintenance and concept drift detection system optimized for the Edge. Combines Arduino/Raspberry Pi hardware logging, OC-SVM anomaly detection, and empirical latencies/energy profiling.

Smart fruit ripening monitor utilizing NICLA Sense and Arduino Nano. Integrates BLE-to-MQTT python routing, embedded machine learning inference, and empirical power consumption telemetry.

Knowledge graph platform mapping the Portuguese DRE corpus into OWL/RDF structures. Includes an analytical Flask REST API, SQLite storage indexers, and a single-page management UI.

End-to-end ML solution for predicting urban traffic congestion levels in the city of Porto, following the CRISP-DM methodology. Achieved a competitive performance that aligns with the top 4 elite positions in generalization capability.

Modular QoS architecture using VNFs, including access control, traffic classification, policing, scheduling, and monitoring. All components were deployed across a virtualized network core using IP-based service chaining.

Constructed a custom language parser pipeline delivering robust operator precedence mapping into targeted custom bytecode VMs.

Interactive Blender simulator handling sophisticated runtime script logic paths for real-time key transformations and sound indexing.

Feedforward binary neural network analyzing 20x20 matrices using early stopping metrics and early split constraints.

Engineered enemy path state machines and responsive vector input threading using complex Java structures.

Structured sandbox collision detection algorithms rendering shifting, spinning polygonal maps.