Two disciplines, one client relationship — strategic research guidance and applied data science, delivered personally by Dr. Delva. Not templated, not delegated.
Over a decade across oncology, molecular biology, and neuroscience — from bench to strategy, for academic teams, clinical groups, and biotech organizations.
Experimental design strategy, hypothesis development, and methods selection — an outside expert perspective before you commit resources to a direction.
Manuscript development and grant proposal strategy (DFG, ERC, NIH) grounded in what reviewers actually look for.
Doctoral & postdoc coaching, fellowship application prep, and career transition support from someone who has navigated all three.
Applied machine learning and statistical analysis — trained through Le Wagon's Data Science & AI program and proven on real biomarker data.
Founder of a dual-engine platform combining iPSC-derived multi-omics profiling with an AI biomarker layer for neurodevelopmental disorder risk stratification. Built and validated a full classification pipeline (PCA, logistic regression, Random Forest) on public transcriptomic data, achieving results competitive with published academic benchmarks.
See the project →EDA, statistical testing, and publication-quality figures — Python/R pipelines built for reproducibility.
Classification & regression, feature engineering, model validation, and biological ML pipelines from raw data to interpretable model.
BI dashboards, GA4/GTM analytics reporting, SQL/BigQuery pipelines, and data storytelling for non-technical stakeholders.
The same tier structure applies whether the work is a research advisory sprint or a full modeling pipeline. Full current pricing lives on Services & Pricing.
Entry-level engagement
Mid-tier engagement
Full-scope engagement
Every case study below was designed, coded, and written personally — the same rigor and format applied to every client engagement.
Sleep Quality & Stress Biomarker Dataset · N = 180 participants. Statistical summary, publication-quality visualizations, and interpretation grounded in current sleep science literature.
Sleep Disorder Risk Classification · N = 300 participants, 7-feature biomarker dataset. Three models built, compared, and validated with biological interpretation.