Scientist I/II, integrated Technology and Exploration (iTX)

insitro•South San Francisco, CA

1d•Onsite

About The Position

At insitro, AI, large-scale biology, and multi-modal target and drug discovery are integrated into one system. The company builds an integrated loop where models generate predictions, experiments generate ground truth, and each cycle updates the system's knowledge. This role is for a Scientist to generate the ground truth that this system learns from. The ideal candidate is grounded in cell biology, treats AI-generated predictions as hypotheses to be tested experimentally, and is motivated by building systems that learn through contact with reality. The Scientist will design and run targeted cell-based experiments using both manual and automated lab workflows to address specific questions from models and programs. When experimental results and model predictions differ, the Scientist is responsible for diagnosing the discrepancy, determining its root cause (model, experiment, or experimental system), and guiding the system's subsequent learning. Success in this role is measured by the system's increased accuracy, better calibration, and enhanced usefulness. This position is based at insitro's South San Francisco headquarters, requiring five days a week onsite presence to facilitate close feedback loops between experiments, automation, and AI-driven analysis. The role reports to the Director of Integrated Technology Exploration.

Requirements

Strong foundation in cell-based experimental biology, with a PhD in a biological science (e.g., cell biology, biochemistry, genetics, or related field) and 2+ years of industry experience, or an MS with equivalent depth of industry experience
Treat AI-generated predictions the same way you treat any hypothesis: worth considering, worth testing, never accepted without experimental evidence
Comfortable working without a playbook. AI-native experimental science is a new discipline, and you're ready to help define what good practice looks like
Design experiments to be maximally informative, not maximally confirmatory. An experiment that cleanly rules something out is as valuable as one that validates it
Comfortable designing and executing experiments in both manual and automated laboratory environments, and you see automation as a tool for learning rather than simply for throughput
Analytically fluent with experimental data. You can work with imaging readouts, plate-level metadata, and gene-level results
Want to understand how models reason, where they fail, and how experiments can make them better
Work well independently and in multidisciplinary settings, bringing clarity, curiosity, and critical thinking to ambiguous problems

Nice To Haves

Experience with high-content imaging or morphological profiling
Familiarity with CRISPR screening (pooled or arrayed formats)
Experience working with laboratory automation or LIMS

Responsibilities

Generate ground truth datasets from cell-based experiments that serve as training and validation data for computational models
Design and execute experiments to test and validate/invalidate AI-generated predictions, selecting cell models, perturbations, readouts, and timepoints based on what the system most needs to learn
Execute experimental work across manual bench workflows and automated platforms, including imaging-based phenotyping, perturbation screens, and multi-modal molecular readouts
Diagnose discrepancies between model predictions and experimental outcomes, determine their root cause, and articulate what the system should learn or change as a result
Maintain structured experimental records where quantitative claims are sourced, unexpected results are documented, and findings are accessible to both human collaborators and computational systems
Apply scientific judgment to decisions about when evidence is sufficient, when uncertainty remains too high, and when approaches should be revised or stopped
Collaborate with biologists, automation engineers, and machine learning and data scientists to translate experimental insights into model improvements and guide subsequent experimental questions

Benefits

401(k) plan with employer matching for contributions
Excellent medical, dental, and vision coverage as well as mental health and well-being support
Open, flexible vacation policy
Paid parental leave of at least 16 weeks to support parents who give birth, and 10 weeks for a new parent (inclusive of birth, adoption, fostering, etc)
Quarterly budget for books and online courses for self-development
Support to attend professional conferences that are meaningful to your career growth and role's responsibilities
New hire stipend for home office setup
Monthly cell phone & internet stipend
Access to free onsite baristas and daily lunch for employees who are either onsite or hybrid
Access to a free commuter bus network that provides transport to and from our South San Francisco HQ from locations all around the Bay Area