Building Biologist Jobs in Washington, DC (NOW HIRING)

ABOUT BULLFROG AI

BullFrog AI (NASDAQ: BFRG) is a computational biology company spun out of the Johns Hopkins Applied Physics Laboratory. We sit at the intersection of AI and drug discovery, building platforms that help pharmaceutical and biotech companies make better, faster decisions across the drug development lifecycle:

• bfPREP™ — biology-aware data harmonization that unlocks the value trapped in fragmented, multi-site clinical and omics datasets
• bfLEAP™ — causal AI analytics for patient subgroup discovery, biomarker identification, and drug target prioritization
• bfARENAS™ — structured multi-criteria decision support for high-stakes portfolio, indication, and go/no-go decisions

THE ROLE

bfARENAS can already rank every human protein by a plain-language functional label, from relevance to dopaminergic tone to inflammasome activation. This role owns the science behind those rankings: what each label means, what evidence stands behind it, and how we show a ranking is sound.

BullFrog AI is seeking a Principal Computational Biologist to lead the science of the functional-ranking system inside bfARENAS. The ranking machinery already exists: give it a plain-language label such as relevance to dopaminergic tone, a specific diabetes symptom, or inflammasome abundance and activation, and it orders all human proteins against that label. Choosing which labels to rank is a shared call with the commercial and product teams. Making sure a label is well-posed, that a ranking means something specific, and that the answer is right, is yours.

The system is early and experimental. We are exploring how these ranked lists could support research and, over time, whether they could become something we offer more widely. Building the evidence behind these rankings, so they stand up to outside scrutiny, is the part you would own. In the first year, the work is mostly to establish how we back a ranking and to set a bar we can defend before anything goes out. This is a senior individual contributor role, and for now a solo one.

WHAT YOU'LL DO

Evidence and Validation

• Own the evidence that establishes a ranking is right, especially where there is no clean answer key
• Decide what each ranking gets compared against: known gene sets, genetic and perturbation evidence, curated associations, orthogonal data, and expert-chosen positives and negatives
• Build the benchmarks and comparisons that back a ranking, and set the bar it must clear before it goes out

What We Rank, and On What Grounds

• Bring the scientific grounding to label design, a shared effort with the commercial and product teams: what a label like dopaminergic tone or inflammasome activation should mean, and whether the system can answer it well
• Turn a loosely worded concept into a precise, answerable question, so a ranking means something specific instead of something vaguely plausible
• Say early when a proposed label is ill-posed or likely to produce a confident wrong list, before compute is spent generating it

Make Rankings Defensible

• Make sure any ranking that leaves the building can be defended to a scientist who did not produce it
• Support the commercial and product teams as they weigh which labels to invest in, grounding those choices in what the system can actually deliver
• Help judge where limited compute is best spent, since generating a label is expensive

What We’re Looking For

Education

• PhD in computational biology, bioinformatics, genetics, systems biology, or a closely related quantitative field

Experience

• 5+ years in a relevant field, with a real track record of designing evaluations for problems that have no clean gold standard, and a nose for which comparisons tell you something and which just look reassuring
• Deep enough in functional genomics or disease biology to tell whether a ranked list of proteins is plausible, with breadth across areas rather than depth in a single disease
• Comfortable with biological concepts that resist a tidy definition, and able to turn one into something you can actually measure and defend to another scientist
• You have run enrichment analyses and GSEA in earnest and come away unconvinced, because you have seen that the annotations feeding them are the largest source of nonsense in the output, second only to the often-ignored null hypotheses underneath
• Fluent enough in Python to take the system’s output apart, build your own checks, and find where it breaks
• Must be legally authorized to work in the United States

Desired Skills

• Familiarity with target-disease association resources such as Open Targets, and a view on where they fall short
• Deep familiarity with the resources that make good ground truth: curated gene sets, genetic evidence, perturbation screens, and expression atlases
• An understanding of how language-model scoring goes wrong, and a working suspicion of output that sounds right
• A publication record in functional genomics, target discovery, or disease biology
• A feel for how target prioritization actually gets used when a team decides what to work on
• Familiarity with disease and phenotype ontologies, and how one concept maps to many terms across databases

WHO WILL THRIVE HERE

• Scientifically rigorous. You distrust a ranking that looks right until you have checked it against something independent
• Comfortable with ambiguity. You can take a vague functional concept and decide, defensibly, what it should mean
• Intellectually versatile. You move easily across disease areas and kinds of evidence, and you like problems that do not sit in one field
• Independently driven. You can scope, run, and judge this work without much oversight
• Collaborative by default. Remote-first is second nature, and label design here is a team sport; you communicate early and close loops

WHAT WE OFFER

Competitive base compensation with eligibility for performance bonus and stock options based on individual and company performance. Full benefits from day one including medical, dental, and vision coverage, short-term disability, and 401(k) enrollment, 15 days of paid time off annually plus 11 paid holidays annually, and maternity and paternity leave.