Applied AI Scientist & Research Pharmacist
Youssef Abo-Dahab
Bridging advanced machine learning and the biomedical domain. Building multimodal deep learning frameworks for drug repurposing, genomic variant interpretation, and target discovery.
UCSF & Stanford University
About
Applied AI Scientist & Research Pharmacist bridging the gap between advanced machine learning and the biomedical domain. I develop multimodal deep learning frameworks for drug repurposing, genomic variant interpretation, and target discovery.
My technical expertise spans Graph ML (PyTorch & PyG), molecular docking, PK/PD modeling, and pharmacogenomics, integrating these distinct disciplines to build rigorous, biologically interpretable models.
I leverage a unique "Clinician-Coder" background (Pharm.D. + MS in AI) to validate computational predictions against real-world biological and clinical data, with training at both UCSF & Stanford University.
Affiliations
UC San Francisco
MS in AI & Computational Drug Discovery
Stanford University
CS224W: Machine Learning with Graphs
Focus Areas
Experience
University of California, San Francisco
Zhao Lab (Capstone Project)
Jul -- Dec 2025Multi-Modal Computational Drug Repurposing: Conducted large-scale benchmarking of computational docking and ML-based scoring approaches. Integrated optimized pipelines with molecular simulations, scRNA analysis, UCSF's clinical data (RWE), and pharmacology-informed ML models to discover drug repurposing candidates for Parkinson's Disease.
Giacomini Lab
Since Feb 2025Developed "TripleVar," a multimodal deep learning framework to predict pathogenic variants in SLC19A2/A3 thiamine transporters by integrating genomic (DNABERT-2) and protein (ESM-2) embeddings with allele frequency data. Achieved 93.3% accuracy and 1.0 AUC on ClinVar benchmarks. Deployed pipeline to screen ~29,000 unclassified variants, identifying 111 high-confidence pathogenic candidates currently undergoing functional HEK cell-based uptake assay validation. Abstract accepted for ASCPT Annual Conference 2026.
Instructor: Digital Therapeutics
Jan -- Feb 2025Developed and taught a 6-session mini-course on Digital Therapeutics for Pharm.D. students.
Benet Lab (Research Assistant)
Sept -- Oct 2024Studied the application of Kirchhoff's Laws to model drug elimination without traditional differential-equation frameworks under Dr. Leslie Z. Benet.
Egyptian Ministry of Health
Head of Health Education Department
May 2023 -- Jul 2024Led health education strategy for Marsa Alam district (5 towns); directed campaigns on disease prevention, reproductive health, and community wellness.
Clinical Pharmacist
Mar 2022 -- May 2023Optimized medication therapy through evidence-based pharmacotherapy and multidisciplinary collaboration.
Primary Care Provider
Dec 2021 -- Mar 2022Examined, diagnosed, and treated patients with non-critical conditions. Led a team of 4 nurses.
Al-Azhar University
Microbiology Research Group
Investigated antimicrobial activity of fungal metabolites and performed in-silico evaluation using molecular docking (AutoDock via PyRx).
Medical Research Writing
Scientific Communication & Evidence Synthesis
Authored 40+ medical review articles across multiple specialties and conducted systematic literature analyses, demonstrating strong skills in scientific communication and evidence synthesis.
Ghamra Military Hospital (Military Service)
Hospital Pharmacy & Logistics
Dispensed 100-400 prescriptions daily; additional duties included logistics and military training.
Sohag University
Medicinal Chemistry Research Group
Designed anti-cancer molecules using ChemDraw and molecular docking, followed by wet-lab synthesis and structure verification via HNMR and C-13 NMR spectroscopy.
Publications
Anti-Staphylococcal, Anti-Candida, and Free-Radical Scavenging Potential of Soil Fungal Metabolites: A Study Supported by Phenolic Characterization and Molecular Docking Analysis.
Al Mousa AA, et al.
Current Issues in Molecular Biology 2024; 46(1):221-243
Wound Healing, Metabolite Profiling, and In Silico Studies of Aspergillus terreus.
Al Mousa AA, et al.
Current Issues in Molecular Biology 2024; 46(10):11681-11699
Unravelling fungal genome editing revolution: pathological and biotechnological application aspects.
Hassane, A.M.A., et al.
Archives of Microbiology 2025; 207, 150
Convolutional Neural Network Models for Binary and Multi-Label Lung Disease Detection from Chest X-Ray Images (BMI212 Coursework).
Abo-Dahab, Y., Oner, S., Jeffrey, C., Motevalli, S., & Abbasi Asl, R.
Zenodo 2025
Entresto (Sacubitril/Valsartan): How one company managed to have a monopoly on the most important drug for heart failure.
Abo-Dahab, Y., & Arzamassky, A.
Zenodo 2025
ORCID: 0009-0007-2539-3049
Projects
Pharmacology Knowledge Graph for Drug Repurposing
94% AUC / 244M ParamsBuilt a multi-modal pharmacology knowledge graph using Graph ML at Stanford. Final model achieved 0.949 AUC with 2M nodes and 244M parameters using schema-expanded TransR. Published user-friendly interfaces on Hugging Face.
Multi-Modal Drug Repurposing for Parkinson's Disease
25,410 patientsCapstone project integrating virtual screening, scRNA-seq, pharmacology knowledge graphs, and real-world evidence from 25,410 UCSF patients. Identified Atorvastatin as lead candidate (HR 0.920; 95% CI: 0.86-0.97).
TripleVar: Multimodal Variant Pathogenicity Predictor
93.3% Accuracy / 1.0 AUCDeveloped a multimodal deep learning framework integrating genomic (DNABERT-2) and protein (ESM-2) embeddings to predict pathogenic variants in thiamine transporters. Screened ~29,000 unclassified variants, identifying 111 high-confidence pathogenic candidates. Abstract accepted for ASCPT 2026.
FDA Molecule Similarity Finder
Tool allowing users to identify the 10 FDA-approved molecules most structurally similar to their compound using Morgan fingerprints. Newer version integrates multiple comparison methods.
AI-Driven hERG Channel Inhibition Predictor
Predicting inhibition of the human hERG channel using AI methods for cardiac safety assessment in drug development.
Lung Disease Detection from Chest X-Rays
CNN models for binary and multi-label lung disease detection from chest X-ray images (BMI212 Coursework). Published on Zenodo.
Population Pharmacokinetic (PopPK) Model
Development of a Population Pharmacokinetic model with covariate analysis using NONMEM for drug dose optimization.
Entresto Drug Discovery Monograph
Comprehensive monograph on Sacubitril/Valsartan: how one company managed a monopoly on the most important drug for heart failure.
Skills
Drug Discovery & Molecular Modeling
Machine Learning & Deep Learning
Programming & Tools
Pharmacometrics & Data
Cloud & Infrastructure
LLM & Agentic AI
Wet Lab
Clinical Practice
Languages
Education & Awards
Education
University of California, San Francisco
San Francisco, California
Master of Science in Artificial Intelligence & Computational Drug Discovery & Development
Stanford University
Stanford, California
CS224W: Machine Learning with Graphs (UCSF-Stanford Exchange Program)
Sohag University
Sohag, Egypt
Doctor of Pharmacy -- Pharm.D. (U.S. Equivalent), Honors
Awards
UCSF AICD3 Merit Scholarship
$45,000
ASCPT 2026 Poster Abstract & Student/Trainee Travel Grant
$500
Certifications
Get in Touch
I'm always interested in discussing new research opportunities, collaborations in computational drug discovery, or the intersection of AI and healthcare. Feel free to reach out.
San Francisco, California, USA