PhD Candidate · UC Riverside

Faisal B. Ashraf

I build structure-aware language models for understanding and designing antibody–antigen interactions. My research integrates protein sequences with 3D structural context to predict binding, capture mutation effects, and guide antibody optimization — working under the supervision of Prof. Stefano Lonardi.

Broadly, I am interested in protein language modeling, structure-based learning, graph neural networks, and generative models for molecular and antibody design. My long-term goal is to build AI systems that bridge machine intelligence with biological insight for real-world biomedical and therapeutic applications.

Faisal Bin Ashraf

Latest News

Dec '25
Sub-reviewer at RECOMB'26
Aug '25
Structure-aware LLM paper published at J. Chem. Inf. Model.
Jun '25
Bioinformatics R&D (Machine Learning) internship at Exact Sciences, San Diego
Mar '25
Presented at AAAI 2025 FMs4Bio workshop
Jun '24
Passed PhD Qualifying Exam
View all news →

Research Experience

2023 — present
Graduate Student Researcher
University of California, Riverside · Computer Science & Engineering
Structure-aware protein language models, contrastive learning for antibody binding, SPICE web platform. Published in JCIM, PLOS, AAAI, NeurIPS workshops.
Jun — Sep 2025
Bioinformatics R&D Intern
Exact Sciences · San Diego, USA
Fine-tuned nucleotide transformer models for probe optimization in hybrid-capture cancer detection workflows.

Education

Ph.D. in Computer Science (expected Dec 2026)
University of California, Riverside
Focus: Machine Learning & Computational Biology. Advisor: Prof. Stefano Lonardi. Passed qualifying exam Jun 2024.
M.Sc. in Computer Science
University of California, Riverside
IUT logo
B.Sc. in Computer Science & Engineering
Islamic University of Technology, Bangladesh

Selected Publications

Full list on Google Scholar ↗  ·  All publications →

J. Chem. Inf. Model. · 2025
Predicting Antibody-Antigen Interactions with Structure-Aware LLMs: Insights from SARS-CoV-2 Variants
DOI: 10.1021/acs.jcim.5c00973 ↗
Structure-aware ESM2 framework achieving 0.93 AUC for binding and 0.95 for neutralization prediction, outperforming DeepAIR, AntiBERTa, AbMap, and A2Binder.
Preprint (bioRxiv)
A Large Language Model Guides the Affinity Maturation of Antibodies Generated by Combinatorial Optimization Algorithms
Faisal Bin Ashraf, Karen Paco, Zihao Zhang, et al.
DOI: 10.1101/2024.12.19.629473 ↗
Ab-Affinity achieves ~160× binding affinity improvements through ML-guided antibody optimization.
Web-based Research Platform · 2025
SPICE: Structural Protein Interaction Complex Evaluator
Faisal Bin Ashraf, Stefano Lonardi
Webapp: spice.cs.ucr.edu ↗
End-to-end web platform for comparative structural analysis of protein–protein complexes.
View all publications →

Contact

Feel free to reach at faisal.b.ashraf@gmail.com or follow me on linkedin.