Member of Technical Staff: Statistical Genetics

Radical Numerics•San Francisco, CA

About The Position

Radical Numerics is an AI research lab building general biological intelligence. The company’s mission is to master the code of life and reduce human suffering. Its work includes developing biological foundation models such as Evo, Evo 2, and Omnii-1, while addressing the responsibility to defend against the potential misuse of generative biology.

Requirements

PhD in statistical genetics, human genetics, computational biology, biostatistics, or a related field, or substantial industry experience working with population-scale genetic data.
Deep working knowledge of statistical genetics concepts and methods, including GWAS, linkage disequilibrium, ancestry and population structure, heritability, fine-mapping, QTL mapping, rare variant analysis, polygenic risk, and variant annotation.
Experience with large genetic resources such as UK Biobank, All of Us, TOPMed, gnomAD, GTEx, FinnGen, ENCODE, or similar datasets.
Strong Python or R programming skills and experience building reliable, reproducible scientific analyses or software.
Experience working with agentic coding tools.
Clear communication skills and ability to bridge scientific context with engineering teams and partner organizations.
Curiosity and resilience when tackling open-ended scientific challenges.

Nice To Haves

Hands-on experience developing or benchmarking polygenic risk score methods.
Familiarity with machine learning for genomics, including sequence models, variant effect predictors, regulatory models, multimodal models, or biological foundation models.
Experience with colocalization, Mendelian randomization, TWAS, causal inference, cross-ancestry genetics, admixed populations, or privacy-preserving genomic analysis.
A track record of building reproducible pipelines, shared resources, open datasets, or benchmarking frameworks used by other scientists.

Responsibilities

Build and evaluate large-scale statistical genetics resources for model training and assessment, including GWAS summary statistics, QTL maps, fine-mapping results, variant annotations, haplotypes, population reference panels, and biobank-scale phenotype data.
Design benchmarks that assess whether models capture genetic architecture, including linkage disequilibrium, ancestry, constraint, polygenicity, pleiotropy, regulatory effects, rare variant burden, and cross-population generalization.
Partner with AI/ML engineers to analyze model behavior on variant effect prediction, disease association, genotype-to-phenotype prediction, and regulatory region interpretation.
Develop practical standards for genetic data provenance, quality control, leakage prevention, population bias assessment, privacy, consent, and responsible use of human genetic data.