Ai Biologist - Peptide Discovery

LatchBio•San Francisco, CA

About The Position

Latch builds scientific benchmarks for AI agents across biology, working with frontier labs and pharmaceutical companies to test agents on real biological workflows. The AI Biologist will serve as the scientific authority on ground truth for the Peptide Discovery team, evaluating whether agents can reason like expert peptide-discovery scientists and make sound decisions from complex selection, biophysical, structural, functional, and developability data. The role involves designing realistic benchmark tasks and datasets spanning linear, macrocyclic, and constrained peptides against protein targets, including protein–protein interactions. The work covers the discovery arc from target tractability through candidate nomination. The AI Biologist will define expert reasoning and write defensible gold-standard answers, curate data from real programs, build Python analysis pipelines, and collaborate with software engineers to turn scientific judgment into agent workflows. The role also tests agents’ ability to identify artifacts, reconcile conflicting evidence, weigh discovery tradeoffs, and communicate calibrated uncertainty. The AI Biologist will interrogate agent reasoning, identify where and why it fails, explain what evidence would change a conclusion, and score confidence appropriately. This is a remote, hybrid, or onsite role, with a preference for onsite work in San Francisco. The position is a 1099 contract (or W-8BEN internationally) with no stated end date. Work is full-time on the employee’s own schedule; most of the team works 40 or more hours per week. Pay is guaranteed during weeks 1–4 while ramping, then paid per certified task. The company states that most people in the position earn $120,000–$180,000 or more, with uncapped upside. Work authorization is limited to U.S. citizens, permanent residents, and OPT holders; STEM OPT extensions and sponsorship are not supported.

Requirements

PhD in peptide chemistry, biochemistry, biophysics, structural biology, medicinal chemistry, or a related field.
Senior, hands-on peptide-discovery experience, including personal analysis of real program data.
Direct experience in at least one of the following: display selection (phage, mRNA, ribosome, RaPID, or DNA-encoded libraries); SPR/BLI or FP binding kinetics; binding-site mapping by structural MS or NMR; or peptide developability such as PAMPA/Caco-2 permeability and proteolytic or serum stability.
Deep expertise in at least one area—display and selection, biophysics, structural mapping, or peptide medicinal chemistry—and working fluency across the others.
Strong command of experimental design, quantitative data analysis in Python or R, and statistics.
Daily use of Claude, Cursor, or similar AI agents, with the judgment to identify when model reasoning is technically correct but scientifically unsound.
Ability to overrule an AI agent when evidence demands it and openness to updating conclusions when shown to be wrong.
Ability to calibrate confidence and communicate uncertainty rather than defaulting to binary calls.
Full-time commitment on the candidate’s own schedule; most of the team works 40 or more hours per week.
Work authorization as a U.S. citizen, permanent resident, or OPT holder; the company cannot support STEM OPT extensions or sponsorship.

Nice To Haves

Peptide-format and chemistry experience, such as macrocyclization, hydrocarbon stapling, N-methylation, D- and non-canonical residues, or HELM encoding.
Target tractability or de novo design experience, such as PPI hotspot analysis, human-genetic evidence for therapeutic direction, or computational design with Rosetta or RFdiffusion.
Leadership experience; the company says people who perform well and bring leadership experience can move quickly into senior roles.

Responsibilities

Define expert reasoning for peptide-discovery questions and write gold-standard answers with supporting rationale.
Design realistic scenarios spanning target tractability, hit generation, hit confirmation and kinetics, binding-site mapping, functional confirmation, and lead optimization.
Curate datasets from real programs, including display-selection read counts, SPR/BLI sensorgrams, fluorescence-polarization titrations, affinity-selection MS, HDX-MS, crosslinking-MS, NMR chemical-shift perturbation, dose-response, permeability, and stability data.
Build Python analysis pipelines for enrichment analysis and kinetic and dose-response fitting, producing verifiable ground truth with tolerances robust across replicates and fitting routines.
Partner with software engineers to translate scientific judgment into new workflows.
Test whether agents distinguish true signal from artifacts such as propagation or expression artifacts, aggregation or superstoichiometric binding, and mistaken interpretation of HDX protection as direct contact.
Adjudicate conflicting evidence, reconcile orthogonal binding-site datasets, distinguish competitive from allosteric engagement, and identify experiments that would resolve uncertainty.
Evaluate whether agents weigh tradeoffs among binding, function, potency, and developability, including the effects of added constraint or affinity on other important properties.
Interrogate agent reasoning step by step and challenge conclusions that would not withstand project review or a lab meeting.
Identify why agent reasoning fails, explain what an expert would do instead, and specify what evidence would change the conclusion.
Score agent confidence, rewarding explicit uncertainty rather than overconfident conclusions on ambiguous data.