Scientist II/ Senior Scientist, BioML

About the role

Lila is building a platform where AI and automation work together to solve the hardest problems in biomedicine. Within Life Sciences AI (LSAI), we are building the loop where AI, automation and experimental biology co‑evolve. We are seeking a Scientist II/ Senior Scientist, BioML to work at the intersection of mechanistic modeling and experimental design. You will help build the biological mechanism layer of that loop: creating reasoning systems that decompose mechanistic hypotheses, judge existing evidence, and identify open questions; encode what good mechanistic adjudications look like; build evaluations that measure a system against those standards; and close the loop by designing experimental campaigns that generate the data needed to ground or test the system. You will own the analysis of the returned data, decide what is worth measuring, and iteratively refine the loop. The ideal candidate brings strong hands‑on experience in computational biology, BioML, or applied ML for biological data and can partner effectively with experimental scientists, ML researchers, and platform engineers to define tractable scientific questions, design rigorous analyses, interpret model outputs in biological context, and contribute to workflows that connect computational predictions with experimental decisions. This is an individual‑contributor role focused on execution, scientific judgment, and cross‑functional collaboration in a fast‑moving, interdisciplinary environment. Lila Sciences is building Scientific Superintelligence™ to solve humankind’s greatest challenges, combining advanced AI models with proprietary AI Science Factory™ instruments into an operating system for science that executes the entire scientific method autonomously.

Responsibilities

• Encode mechanistic reasoning into systems and build the evaluations that hold them to those standards.
• Translate how a good scientist decomposes a hypothesis — what must be established, in what order, what distinguishes live alternatives — into structured frameworks and reasoning workflows that operate autonomously.
• Define what constitutes a good answer and score whether a system reached it for the right reason rather than a plausible‑sounding one.
• Analyze biological data that grounds these systems: hands‑on work with single‑cell, perturbation, proteomics imaging, and genetic data; build pipelines that turn raw measurement into model‑ready evidence; run quality control and annotation‑stability analyses; produce quantitative results the reasoning layer depends on.
• Design experimental campaigns and close the loop: specify what to measure, in which contexts, at what precision and scale — both to anchor systems in real data and to generate outcome‑verified cases for training and scoring; analyze returned data, update assessments, and articulate what changed.
• Build outcome‑verifiable benchmarks and set evidence standards, including retrodiction sets that score a system on predicting documented outcomes from pre‑decision evidence only, with the evidence boundary enforced.
• Define what counts as direct versus inferred evidence and when a computational prediction may substitute for a measurement.