Data Scientist
- Developed machine-learning classifiers for aircraft flight-phase identification using operational telemetry, achieving an F1 score of 0.98 on an evaluation set
- Designed and validated FFT-based audio signal-processing pipelines to extract frequency-domain features, estimate aircraft engine RPM, and determine engine operating states, achieving R2 = 0.80 against ground-truth RPM measurements
- Processed 1,500 telemetry/audio samples, with a reported 30% processing-speed improvement and 80% less manual data preparation
- Developed automated telemetry analytics, REST APIs, and a full-stack monitoring dashboard supporting over 200 users
Flight data monitoring case study →
Machine Learning Engineer
- Developed predictive models for vending-machine delivery-failure forecasting using 5 million+ transaction records across approximately 150 machines, with a reported F1 score of 0.85
- Built automated ETL, data-processing, and operational analytics workflows, with reported 85% faster reporting
- Developed KPI-monitoring workflows covering over 1 million transaction records to support operational performance analysis and decision-making
Operational Data Analyst
- Optimised SQL queries and automated business-intelligence reporting, with reported 45% faster query execution and a 75% reduction in reporting effort
- Developed automated, customised CRM platforms for customers using SQL and Omniaplace, incorporating tailored functionality to improve performance, operational efficiency, and versatility
- Designed reproducible data-processing and reporting workflows to support operational decision-making
Digital Engineering Intern
- Developed computational-geometry algorithms and optimised data structures for tire cross-section simulation and finite-element meshing workflows
- Achieved a reported 20% computational speed improvement through algorithmic and data-structure optimisations
- Implemented a closest-point utility for cubic Bézier curves to support geometry projection and meshing operations
Machine Learning Intern
- Integrated an InsetGAN-based whole-body image-generation pipeline (StyleGAN-Human with InsetGAN joint optimisation) with reported evaluation on 500 generated samples
- Developed and deployed model inference through FastAPI for integration with a Swift iOS application, with a reported 2.3-second inference latency
Audio engine monitoring at Intuos
At Intuos Srl, I developed an aircraft engine audio-monitoring prototype using Raspberry Pi, FFT-based signal processing, and an ESP32 display. The audio-monitoring system complements the recorded-telemetry dashboard as a separate engineering workflow.