I build gene therapies the way I build software: from first principles, source open, no ceremony.

BUILDING

AAV gene-replacement therapy for DYT-SGCE myoclonus-dystonia at Arcivus Lab.

STUDYING

Bioinformatics @ Langara. UBC Microbiology and Immunology, 2027 intake.

OPEN TO

Wet-lab partners for AAV production. Bioinformatics consulting (Python / LLM pipelines).

Patient-specific molecular identity should drive therapy design, not population averages.

The concrete test case is a single frameshift in the SGCE gene¹ — one nucleotide duplicated, one premature stop codon, one truncated ε-sarcoglycan. Arcivus Lab's first program is an AAV vector that replaces the broken copy. No small molecules. No symptom management.

NorthProt is the silicon half: Python, Claude API, MCP pipelines, pgvector retrieval over the primary literature. Arcivus is the carbon half: capsid engineering, construct design, wet-lab validation. The same person runs both sides because the target-identification-to-construct gap² is where rare-disease programs die, and collapsing it is a tooling problem before it is a biology problem.

From first principles, source open, no ceremony³ — this is how the work is done, not how the site is marketed. Every tool Arcivus builds ships open-source. Every decision is traceable to a paper or a sequence. Small. Real.