AI Training through Scientific Task Recording
Budget: ₹1,500 – ₹12,500 INR
Paid Project for Life-Sciences Experts — Project
Brief
Recording real analysis work to help train AI. Remote, flexible, and paid per task.
What this is
This project collects recordings of experienced scientists performing real analysis work in
scientific software. As each task is completed in the contributor's usual tool, the screen, actions,
and spoken reasoning are captured. These recordings are used to train AI to operate scientific
software properly, a capability current models still lack.
In short: the contributor's expertise doing the work is the product. No surveys and no
clicking exercises, only real, careful analysis of the kind already performed in the contributor's
normal work.
Which tools
Experts are being recruited in one or more of these five tools:
• RStudio (R) - bioinformatics, statistics, figures (RNA-seq, tidyverse, Seurat, etc.)
• GraphPad Prism - statistics and publication graphs
• ImageJ / Fiji - scientific image analysis and quantification
• SAS - clinical-trial / biostatistics programming
• SPSS - medical, public-health and survey statistics
Workflows using a single tool are acceptable if realistic and domain-relevant, but
cross-application workflows are preferred, as they better reflect how multi-tool experts work
in practice.
Brief
Recording real analysis work to help train AI. Remote, flexible, and paid per task.
What this is
This project collects recordings of experienced scientists performing real analysis work in
scientific software. As each task is completed in the contributor's usual tool, the screen, actions,
and spoken reasoning are captured. These recordings are used to train AI to operate scientific
software properly, a capability current models still lack.
In short: the contributor's expertise doing the work is the product. No surveys and no
clicking exercises, only real, careful analysis of the kind already performed in the contributor's
normal work.
Which tools
Experts are being recruited in one or more of these five tools:
• RStudio (R) - bioinformatics, statistics, figures (RNA-seq, tidyverse, Seurat, etc.)
• GraphPad Prism - statistics and publication graphs
• ImageJ / Fiji - scientific image analysis and quantification
• SAS - clinical-trial / biostatistics programming
• SPSS - medical, public-health and survey statistics
Workflows using a single tool are acceptable if realistic and domain-relevant, but
cross-application workflows are preferred, as they better reflect how multi-tool experts work
in practice.