Yu S. Huang

- Define and execute long-term technical strategy for AI-driven precision oncology, aligned with corporate product pipelines and business goals. - Lead the development of multimodal AI platforms integrating sequence, structure, and epigenomic data for non-invasive cancer detection. - Built enterprise-grade AI computing infrastructure (K8s, PyTorch, distributed storage, high-speed interconnect) to support large-scale computing. - Lead and mentor a high-performance team of algorithm scientists, bioinformaticians, and software engineers to deliver end-to-end solutions from in silico modeling to experimental validation. - Led cross-disciplinary team management and promoted tight integration between computational models and experimental biology. - External scientific engagement, conference presentations, high-impact publications, and IP strategy; drove research-to-product translation. - Optimize core bioinformatics algorithms using Deep/Machine/Statistical Learning techniques. - AI models for fixed-panel and custom-panel MRD (Minimal Residual Disease). - Teach Bayesian Statistics, Machine/Deep Learning, Julia/Rust Programming.

- Led the establishment of AI-driven computational biology and drug discovery center and built a mature structure-based drug design & virtual screening system. - Developed Fergie (VAE-based small molecule generation) and Deffini (structure-based virtual screening DNN) to enable structure-guided drug design at scale, with one molecule entering the PCC phase. - Developed core algorithms for genomic variant calling, copy number analysis, and methylation sequencing to support early-stage innovative drug R&D. - Directed national/provincial research projects, built academic-industry partnerships, and delivered high-impact publications. - Taught "Artificial Intelligence" & "Pattern Recognition and Machine Learning". - Taught Julia programming language, Matrix Computations, Optimization. - 2015 CAS Hundred-Talent Program, 2016 China Thousand-Talent Program. - http://www.yfish.org/

- Developed algorithms and pipelines for high-throughput sequencing data analysis. - Built MethylSeq analysis tool on Illumina BaseSpace for bisulfite sequencing data processing. - Developed UFlow, a Directed-Acyclic-Graph workflow system that speeds up Illumina bioinformatics workflows by >50X. - Developed a bioinformatics library in GOlang that sped up some analysis by >100X. - Forensics, cancer, whole-genome, exome competitive analyses.

1. Computational methods for pedigree genomics. 2. NGS parallel workflow and SQL database for ~1000 Vervet monkey genomes project.

1. Conditional linear mixed model for GWAS 2. Graph-theory based algorithms 3. Statistical model for pathway enrichment in GWAS hits 4. Parallel computing pipeline and SQL database for processing large genomic data