André Ferreira Santos

Data Scientist · Biomedical Engineer · Health Informatics Specialist · AI/LLM Explorer

Coimbra, Portugal · afsantos091@gmail.com
GitHub ·  LinkedIn ·  Portfolio

Professional Summary

Data Scientist and Biomedical Engineer with 10 years of experience transforming complex biomedical data into intuitive, real-world solutions. Combines deep technical mastery (Python, R, Neo4j, SQL, ASP.NET) with a human-centric approach to design and usability.

PhD in Biomedical Engineering (University of Coimbra, awarded June 2026 — Aprovado com Distinção e Louvor) on graph analytics & ML in autism genetics, reaching 88.6% accuracy classifying ASD vs. Developmental Delay — 8 points above DSM-5 diagnostic sensitivity for children under 7. Now expanding into LLM and Generative AI applications. Passionate about collaboration, knowledge-sharing, and innovation that simplifies people’s lives.

Highlights & Impact

Technical Profile

Data Science & Analytics
PythonR scikit‑learnPandas NumPySciPynetworkX Altairggplot2EDA Feature EngineeringGraph ML (Neo4j GDS) Machine Learning PipelinesStatistics VisualizationTableuGephi
Software & Development
JavaScriptNode.js JavaDjango Ruby (Rails)Elixir (Phoenix) C#ASP.NET MVCREST APIs GitCI/CD
Infrastructure & Ops
PostgreSQLMySQL Neo4j LinuxShell Script ETL PipelinesData Modeling GitHubPostmanXMLJSON
AI & LLM
Claude CodePrompt Engineering LLM-Assisted Research WikisCLAUDE.md Instruction Design

Professional Experience

Independent Researcher — AI-Assisted Knowledge Systems (LLM Wiki / Forge)
2025 – Present | Generative AI · Knowledge Management · Claude Code
  • Designed a repeatable ingest → discuss → write → link → index → log workflow that turns raw source material into a persistent, cross-referenced markdown wiki using Claude Code, governed by standing CLAUDE.md instruction sets.
  • Built and maintained three instances: a 20+ page PhD defence wiki, a 19-page EU competition (EPSO AD7) exam-prep wiki, and Forge, a 26-page meta-wiki cross-comparing four AI-assisted development methodologies.
  • Converted synthesized knowledge into reusable CLAUDE.md templates and project scaffolds for downstream projects, scaling the method from single-domain capture to comparative, cross-framework synthesis.
PhD Researcher — University of Coimbra · CIBIT & ICNAS
2019 – 2026 (Awarded) | Biomedical Engineering · Neurogenomics · Graph Data Science
  • Applied graph analytics and machine learning to uncover dopaminergic and serotonergic mechanisms differentiating Autism Spectrum Disorder (ASD) and Developmental Delay (DD).
  • Designed and maintained relational (MySQL) and graph (Neo4j) databases integrating genomic CNV data, clinical phenotypes, and ontology terms (GO).
  • Developed ETL pipelines (Python · R · Ruby) for data normalization, annotation (Ensembl Biomart), and integration of >68 000 patient‑CNV relations.
  • Performed complex network analysis (hub detection, modularity, centrality metrics) and trained Random Forest classifiers achieving 88.6% accuracy in differential diagnosis tasks — 8 points above DSM-5 diagnostic sensitivity (78%) for children under 7.
  • Implemented graph‑native ML pipelines (node2vec embeddings, Neo4j GDS) to classify participants via genetic and functional similarity networks.
  • Published peer-reviewed work in the Journal of Personalized Medicine (2022) and Scientific Reports (2026) on serotonin CNV networks in ASD/DD.
  • Collaborated with a research psychologist (CIBIT) to build COVMind, a Django-based remote CBT/mindfulness platform delivering structured therapeutic sessions during the COVID-19 pandemic.
  • Released datasets and code as open‑science resources → Dataverse Repository.
Health Informatics Specialist — Centro Hospitalar Universitário de Coimbra (CHUC)
2017 – 2024 | Clinical Informatics Department · Digital Health & Interoperability Engineering
  • Led the design and deployment of a structured clinical documentation system at CHUC’s Pediatric Neurodevelopment and Autism Unit (UNDA), integrating HL7 CDA and SNOMED‑CT for semantic and structural interoperability.
  • Developed a three‑layer architecture (data access · logic · UI) connected to national EHR systems (SClínico and SONHO V2) via HL7 messaging and REST services.
  • Created a web‑based ASP.NET MVC application allowing clinicians to define reusable templates for structured assessments (e.g., Griffiths Mental Developmental Scale).
  • Implemented SNOMED‑CT terminology server integration for real‑time concept lookup and validation, ensuring high‑quality coded data.
  • Engineered HL7 v2 interoperability modules (via LIGHt gateway) to exchange structured data with the Portuguese National EHR (SNS), enhancing workflow efficiency.
  • Validated and adopted in clinical practice with multidisciplinary teams of psychologists, pediatricians, and data scientists, impacting a potential user base of 65 000+ healthcare professionals nationwide.
  • Delivered measurable improvements in data quality, semantic consistency, and research integration, enabling longitudinal analysis and population queries.
  • Published peer‑reviewed work in the Healthcare (2023) journal.
Research Fellow — CIMAGO (Biomedical Research Unit, University of Coimbra)
2015 – 2016 | Genomic Data Science · Machine Learning · Bioinformatics
  • Collaborated with the Cytogenetics and Genomics Laboratory on cross‑disciplinary projects integrating genomic, epigenetic and clinical data for cancer and neurogenetic research.
  • Developed a LAMP‑based genetic data integration portal to upload and parse Array‑CGH CNV data, standardizing patient records through a secure web interface with gene and chromosome‑level visualization.
  • Replaced manual spreadsheet workflows with a MySQL relational repository linked to UCSC Genome Browser and GeneCards, improving traceability and re‑use of clinical data.
  • Implemented R / Python / Ruby pipelines for feature selection and classification of oral squamous cell carcinoma (OSCC) samples using Random Forest and clustering models.
  • Identified genomic patterns and candidate biomarkers, enhancing reproducibility and scientific transparency.
  • Co‑authored peer‑reviewed articles including Ribeiro et al. (2017) Molecular Cytogenetics and Ribeiro et al. (2016) Eur J Hum Genet (Q1 IF 4.58).
Researcher / Developer — HNotes (Semantic Health Data Capture Platform)
2023 | Independent Exploration · Semantic Modeling & Web Engineering
  • Designed and developed HNotes — a web‑based proof of concept for structured and semantically interoperable clinical documentation.
  • Implemented in Elixir / Phoenix LiveView with Ecto for data persistence, modeling records through SNOMED CT terminology.
  • Defined a six‑field semantic schema (Component, Property, Time, System, Scale, Method) ensuring unambiguous encoding of every clinical observation.
  • Created dynamic real‑time interfaces for data entry and hierarchical organization (notebooks → notes → registries).
  • Demonstrated interoperability using cardiovascular data (blood‑pressure records) with SNOMED‑compliant attributes.
  • Showcased integration of semantic interoperability, functional programming, and interactive web technologies.
  • Explore the live demo: HNotes prototype ▶️

Education

Certifications & Professional Development

Core Strengths

Motivation

I believe technology should be both beautiful and useful. My goal is to apply data science and AI to create digital products that make life simpler and more intuitive.