Howdy, I’m Kelechi Fletcher.

Proof points

30-point recall improvement: For a genomic evidence-extraction workflow, I researched and developed agent-driven schema and prompt optimization. The resulting system improved variant-mention recognition recall by 30 percentage points over the internal baseline.
50,000+ samples / 5+ TB: My bioinformatics experience spans both high-volume sequencing and gene expression analysis. I modernized analysis pipelines and automated reporting across more than 50,000 samples and 5 TB of scientific data, and separately implemented pipelines for gene expression profiling and differential gene expression analysis deployed across cloud and on-premises infrastructure.
Shared foundations across departments: When teams needed more consistent access to critical data and capabilities, I authored shared SDKs and led the development of Dagster assets and orchestrated workflows used across departments.

I turn ambiguous scientific, operational, and data-intensive workflows into software people can understand, trust, and depend on.