Fernando Delgado Chaves

Co-founded an agentic bioinformatics platform delivering executable scientific analyses, figures and reports. Principal author of the agent orchestration layer, an 810-skill scientific skill catalogue, the CLI and E2E tests. Co-built the BixBench evaluation harness: 81.0% on the full public set, 92.0% on Verified-50. Led sales and customer demos; 450+ users, paying biotech customers, ~€234K non-dilutive funding.

CoSy.Bio institute. Led a group on generative AI for biomedical research; secured ~€252K funding (Horizon Europe FLAME); taught a graduate course. - Built LLM agents for literature screening, evidence extraction and research workflows; benchmarked 18 LLMs vs. expert systematic reviews. - REPO4EU network-medicine strategies for host-directed antiviral drug repurposing. - Developed DRaCOoN (differential networks) and omics biclustering for patient stratification.

Developed gene-network reconstruction algorithms (ensemble-and-greedy approach in R/Java) to identify co-expression relationships; applied them to molecular responses to DNA repair and SARS-CoV-2 infection. Streamlined NGS workflows and mentored thesis students. Three research stays building wet-lab skills (CRISPR-Cas9, confocal microscopy, immunostaining).

Thesis: new algorithms for gene network reconstruction.

MSc Advanced Bioinformatic Analysis (9.14/10); MSc Health Biotechnology, Cell Therapy (9.56/10); MSc Higher Education, teaching qualification with CLIL (8.81/10).

Erasmus year at Ruhr University Bochum, Germany.