AI Task Architect for Real-World STEM & Biology

Planet Pharma; San Francisco (CA); Remote; USD 120,000 - 180,000; Full time

Job summary

Planet Pharma (remote) seeks experienced scientists and engineers to design and evaluate frontier AI models using real test, measurement, and process data. You will craft realistic tasks from your practice, run them through AI agents, and assess outputs against professional standards.

Strong coding and scientific computing are required for some projects, with long-form, self-directed work in a fully remote setting.

Qualifications

• 2+ years of applied experience in core sciences or engineering or data science.
• In progress Bachelor’s degree or higher; 2+ years outside undergrad preferred.
• Comfort writing and verifying work in code (Python or similar) and using command line.
• Hands-on with real data and analysis tools (MATLAB, Python, R; SPICE, PCB tools; FEA/CFD; GIS).
• Strong understanding of statistical experimentation and inference.

Responsibilities

• Design challenging, realistic technical tasks drawn from your day-to-day work.
• Run tasks through frontier AI models and evaluate deliverables against professional standards.
• Compare two model outputs and document which performed better and why.
• Write detailed grading rubrics and explain pass/fail criteria.
• Flag concrete failures with evidence: unit errors, misread data, or boundary issues.
• Review and refine tasks built by other experts across disciplines.

Skills

Applied experience

Python

MATLAB

R

Data science

Experiment design

Statistics

English writing

Education

Bachelor’s degree or higher

Tools

MATLAB

Python

SPICE

PCB tools

FEA

CFD

GIS

Chromatography software

Job description

Planet Pharma (remote) seeks experienced scientists and engineers to design and evaluate frontier AI models using real test, measurement, and process data. You will craft realistic tasks from your practice, run them through AI agents, and assess outputs against professional standards.

Strong coding and scientific computing are required for some projects, with long-form, self-directed work in a fully remote setting.