Stefano Siccardi
I specialize in transforming complex analytical and technical challenges into practical solutions, prototypes, and validated project specifications.
My background spans applied research, data modeling, machine learning, NLP, graph analytics, scientific computing, and algorithm design. Over the years I have worked on projects involving medical imaging and radiomics, graph-based machine learning, natural language processing, optimization algorithms, energy modeling, and advanced computational methods.
Rather than focusing on a single technology stack, I specialize in understanding problems, evaluating possible approaches, and rapidly building proofs of concept that validate ideas before larger investments are made.
My typical contributions include:
• Requirements discovery and technical analysis
• Feasibility assessment and solution design
• Data modeling and machine learning prototypes
• Algorithm design and performance analysis
• Technical specifications for development teams
• Troubleshooting and debugging of complex systems
• Collaboration between stakeholders, researchers, and developers
I am comfortable working across disciplines and entering unfamiliar domains when a project requires it. I regularly use AI-assisted development tools to accelerate coding, experimentation, testing, documentation, and technical exploration.
Currently available for part-time engagements (20-24 hours/week), particularly short- to medium-term projects involving data-intensive applications, machine learning, scientific computing, technical discovery, and proof-of-concept development.
REVIEWS FROM CLIENTS
- Developed and validated machine-learning workflows for radiomics and biomedical datasets, including feature extraction, dimension...
- Developed and validated machine-learning workflows for radiomics and biomedical datasets, including feature extraction, dimensionality reduction, model evaluation, and reproducible analytical pipelines.
- Designed NLP and information-extraction methods to transform unstructured legal documents into structured event representations, integrating multiple information sources and validating consistency under conflicting evidence.
- Built quantitative models and computational experiments in Python for scientific research, applying statistical analysis, optimization methods, and numerical simulations to complex real-world problems.
- Investigated patterns in large datasets through feature engineering, hypothesis testing, and model validation, emphasizing reproducibility, uncertainty assessment, and failure-mode analysis.