Computational scientist building reliable research software at the intersection of bioinformatics, scientific machine learning, and systems engineering. I work primarily in Python, Rust, Bash, and SQL, with domain experience spanning influenza virology, viral genomics, drug discovery, and omics analysis.
My work emphasizes reproducibility, explicit data contracts, leakage-resistant validation, deterministic behavior where appropriate, realistic failure testing, and clear separation between software performance and scientific claims.
I am also the founder of SignalForge Advisors and a science-fiction author interested in complex technological systems and their human consequences.
| Project | What it does |
|---|---|
| FluFormer | Scientific machine learning for influenza proteomes: multitask classification, temporal conditioning, and reproducible held-out evaluation |
| SigTrellis | Replicate-aware elastic-net discovery and validation of candidate transcriptomic signatures in bulk and single-cell RNA-seq |
| Virustic2 | Deterministic, quality-aware de Bruijn viral unitig assembly in Rust |
| FluTrees | Scientist-facing influenza HA mutation decision trees with desktop GUI, reports, and reproducible CLI workflows |
| BayesTrellis | Bayesian candidate-signature discovery for bulk and single-cell transcriptomics; currently an explicitly labeled research alpha |
| Antigenic Audit | Detects entity leakage, reversed-role leakage, and temporal overclaiming in influenza antigenicity evaluations |
| Stabilomics | Robust LAD-LASSO stability selection for high-dimensional scientific feature tables |
| RefAudit | Audits downstream sensitivity to reference-genome choice |
| SeqSketch | Deterministic MinHash/LSH sequence indexing, exact reranking, and similarity graphs |
My public projects emphasize strict input contracts, automated testing, reproducible examples, explicit provenance, auditable outputs, and scientific limitations that are stated rather than hidden. I prefer tools whose important assumptions can be inspected and whose failure modes can be tested.
Bioinformatics & computational biology
Influenza · viral genomics · sequence analysis · genome assembly · mutation analysis · transcriptomics · omics
Machine learning & statistics
Protein representations · temporal validation · transformers · sparse modeling · Bayesian modeling · stability selection · leakage auditing
Scientific software engineering
Python · Rust · Bash · SQL · reproducible pipelines · HPC · data provenance · deterministic tooling
- Professional website
- SignalForge Advisors — independent scientific and technical consulting
- Science-fiction projects including SOFTMAX, The Orchard, and The Gospel According to Greg