// biomedical ai specialist · genova, italy
PhD Biomolecular Scientist turned AI Engineer — bridging wet-lab research, structural bioinformatics, and production AI systems for life sciences and code intelligence.
// 01 — about
I'm a Biomedical AI Specialist based in Genova, Italy, with a PhD in Biomolecular and Health Sciences from the University of Urbino Carlo Bo. My career spans wet-lab experimental research, structural bioinformatics, and — more recently — production AI engineering.
At an AI startup in Pisa, I design and engineer large-scale benchmark pipelines for evaluating AI coding agents. My work includes dataset extraction at scale on AWS EC2/S3, LLM-as-Judge evaluation frameworks with HITL calibration, and deep graph and code-based EDA across tens of thousands of bug instances.
My PhD research at Urbino investigated the role of nutraceutical compounds in NAFLD prevention — combining computational phytochemistry, molecular docking, and cellular biology experiments.
I'm open to European research and industry roles at the intersection of AI, bioinformatics, drug discovery, and precision medicine.
// 02 — projects
Designed and engineered a multi-step production-readiness benchmark for evaluating AI coding agents. Built the full pipeline: LLM-as-Judge evaluation with threshold calibration, HITL review, and incident injection across 6,287+ normalised tasks.
Engineered extraction and normalisation pipelines across seven major Python bug datasets: SWE-bench Full, PyResBugs, BugsInPy, TSSM, SWE-Gym Raw, PyTraceBugs, and SWE-smith. Performed graph and code-based EDA mapping ~52,700+ instances against a 12-family, 36-type bug taxonomy. Built tiktoken cl100k_base token analysis pipelines across multi-field code corpora.
Doctoral investigation into the hepatoprotective role of nutraceutical compounds in Non-Alcoholic Fatty Liver Disease (NAFLD). Combined cellular experimentation — mammalian cell culture, fluorescence and electron microscopy, Western blotting — with computational approaches to characterise compound efficacy at molecular level.
Computational analysis of Nigella sativa-derived phytochemicals as inhibitors of Beak and Feather Disease Virus (BFDV). Performed multi-ligand molecular docking with AutoDock Vina, protein structure homology modelling via SWISS-MODEL and MODELLER, and structural visualisation with PyMOL and Chimera UCSF.
// 03 — research
Manuscript in preparation on methodologies for evaluating production-readiness of AI coding agents — covering benchmark design and LLM-as-Judge evaluation frameworks.
Manuscript in preparation drawing on doctoral research in biomolecular sciences and computational approaches to disease mechanism characterisation.
University of Urbino Carlo Bo. Experimental and computational investigation of nutraceutical hepatoprotection mechanisms in non-alcoholic fatty liver disease.
University of Management and Technology, Lahore. In silico screening and molecular docking of plant-derived compounds as candidate antiviral agents against Beak and Feather Disease Virus.
// 04 — contact
Open to European research collaborations, biomedical AI roles, and industry opportunities at the intersection of life sciences and AI.
tayyabanaseem94@gmail.com