Cambridge, Massachusetts, United States
Passionate about driving innovation in therapeutics startups, with a focus on bioinformatics that has directly contributed to an FDA-approved drug and two IPOs. As a former Head of Bioinformatics, I specialized in developing CRISPR-engineered B cell-based therapies for targeted protein secretion, addressing critical medical needs in areas like immunology and oncology. My background includes pioneering work in oncology drug development, utilizing synthetic lethality and advanced analytics for genome-wide CRISPR screens and cancer data mining. This experience has also involved co-developing a whole-genome CRISPR screen algorithm, which has led to promising clinical results, including an ATR inhibitor now in Phase 1 trials. At Google, I expanded the reach of the featured snippets feature, an integral part of Google's search experience that delivers concise answers at the top of search results. My contributions doubled snippet usage in India, showcasing my expertise in C++ and Python for production software on a global scale. In addition, I’m the principal developer of MolBioLib, a C++11 framework designed to accelerate bioinformatics tool development. As a foundational tool, it has supported analysis pipelines in over 15 publications and has been a core component in three startups for their primary analytical pipelines. To expand its functionality, I integrated Perl and R, enhancing the framework's adaptability and enabling more robust bioinformatics solutions. My expertise spans a range of high-impact bioinformatics applications, including microarray, Sanger, and next-generation sequencing, with deep experience in analyzing and integrating complex data types such as RNA-Seq, RIP-Seq, ChIP-Seq, chromosomal conformation, and bacterial 16S regions. With a background in large-scale production software, massively parallel systems, and nonlinear PDEs, I’m skilled in translating complex data into actionable insights. Specialties: C++, Perl, R, bioinformatics, algorithms, natural language processing, microbiome, CRISPR, synthetic lethality, numerical analysis, finite element methods, and large-scale cloud computing environments.
• Bioinformatics consulting company founder. • Developed a locally deployed LLM RAG system using LangChain for analysis of anonymized medical notes. Milvus was used as the vector database. • scRNA-seq analyses of cohorts of cytopenia using nf-core (Nextflow) with custom downstream analyses.
• Led CRISPR guide design analyses for novel cellular therapies, culminating in the selection of a guide currently used in clinical trials for rare diseases. • Designed and implemented a novel amplicon-based CRISPR on-target analysis and comprehensive off-target discovery pipeline, enabling FDA IND clearance for a B cell-based hemophilia B therapy. • Developed and deployed multiple bioinformatics pipelines, including custom-built workflows for amplicon sequence analysis and production-ready pipelines using Nextflow (e.g., nf-core/rnaseq). • Performed orthogonal off-target analyses using Oxford Nanopore Technologies (ONT) long-read sequencing data. • Architected scalable computing infrastructure and developed a multi-omics SQL database on Google Cloud, forming the backbone of Be Biopharma’s bioinformatics platform. • Deployed a GPU-accelerated version of AlphaFold2 on Google Cloud, improving the accuracy of internal protein structure predictions by approximately 50%. • Built and led a high-performing team to analyze preclinical data using rigorous statistical and computational approaches.
• Developed a groundbreaking whole-genome CRISPR screen analysis machine learning (ML) algorithm, leading to two drug discoveries in Phase 1 and Phase 2 trials. • Pioneered an analytic approach for identifying and ranking oncologic targets using synthetic lethality, driving therapeutic advancements. • Established a company-wide multi-omic research database with a user-friendly Shiny app interface on Google Cloud. • Developed programs, such as those parsing mass spectrometry PDFs using optical character recognition (OCR) due to a lack of embedded text, that increased efficiency of scientists by over 500%. Other programs reformat automatically generated Excel spreadsheets to a more human-readable format.
• Led natural language processing (NLP) and ML - including TensorFlow - strategies to boost featured snippets, the product area of Google Search for extracting and summarizing the most relevant search results for users. • Added 13 million additional triggers per day. • Doubled usage in India.