Show me the repetitive work your scientists are doing, and I'll see whether we can automate it.
Tor Brager-Larsen, FounderPelushka's Lab builds practical software and automation for repetitive laboratory workflows, from image analysis and scientific data processing to reporting, bioinformatics, and internal laboratory tools.
Watch Our IntroductionSee How We Work
From scientific workflow to bottleneck to practical automation.
We start with the workflow, find the repetitive work, and build the software to automate it.
Show Us Your WorkflowWhat are your scientists doing manually?
Laboratories generate sophisticated data. The work surrounding that data is often much less automated. We look for the repetitive steps that consume scientific time and determine whether they can be turned into software.
If I show you my workflow, what could you build?
These are examples of the software we build. The company isn't limited to this list. The umbrella is simpler: show us the workflow.
Image analysis
Automate counting, classification, scoring, and quantification from laboratory images.
- Cell counting
- Object detection
- Assay image scoring
- Microscopy quantification
Image-processing scripts, batch pipelines, and measurement exports.
Data automation
Clean, transform, validate, and structure scientific data from existing laboratory software and instrument exports.
- CSV and Excel cleanup
- Schema normalization
- Data validation
- QC flags
Automated pipelines that turn raw exports into analysis-ready datasets.
Reporting
Turn recurring calculations, QC checks, charts, and formatting into automated reporting workflows.
- Standardized templates
- QC summaries
- Charts
- PDF and Excel outputs
Report generators that run on a schedule or on demand.
Bioinformatics
Build reproducible pipelines and custom analysis workflows for sequencing and molecular data.
- RNA-seq analysis
- qPCR data analysis
- Sequence analysis
- Data QC
Reproducible pipelines built with tools like R, Python, and common bioinformatics software.
Laboratory tools
Build lightweight internal applications for samples, experiments, dashboards, data entry, and workflow management.
- Sample trackers
- Experiment dashboards
- Internal databases
- Data-entry tools
Small applications that fit the workflow you already have.
Scientific knowledge tools
Make protocols, SOPs, and internal documentation easier to search and use.
- SOP search
- Protocol Q&A
- Source references
- Human verification
Searchable knowledge systems over your own documents.
What automation could look like
Illustrative examples of the kind of software we build. These are example workflows, not completed client projects.
Example workflow / Illustrative use caseManual image analysis
Scientist reviews hundreds of images and records measurements manually.
Software processes the image set and extracts the relevant measurements.
Structured results ready for review and analysis.
Instrument export to report
Export data, clean spreadsheet, calculate results, make graphs, format report.
Data processing, validation, calculations, QC, report generation.
A repeatable workflow producing standardized output.
Recurring bioinformatics
Analyst manually repeats the same analysis for each dataset.
Standardized reproducible pipeline.
Consistent analysis, QC, and output.
Laboratory SOP search
Search shared folders and PDFs for the correct procedure.
A searchable system over the laboratory's own documents.
Ask a question and retrieve the relevant source material.
Show us the workflow. We'll find the repetition.
You don't need a specification. Bring the problem, and we'll identify where automation fits.
Show us the workflow
We learn how your team actually works today.
Find the repetition
We identify manual, repetitive, error-prone, or inefficient steps.
Build the automation
We create software around the workflow you already use.
Validate with your team
Scientists review the output and determine whether it works in practice.
Deploy and expand
The solution becomes part of the workflow, and we identify further opportunities.
Automation handles repetition. Scientists keep the judgment.
We start with the workflow you already have, use the simplest reliable technology, and leave scientific decisions with your team.
Work with what you already have
We improve existing laboratory workflows rather than requiring wholesale replacement of systems.
Start with the bottleneck
We target specific repetitive work instead of selling unnecessary software.
Use the simplest technology
We choose tools based on the problem, and use AI only where it provides a practical advantage.
Build practically
Start with one workflow. Prove the value. Expand from there.
- Detect
- Organize
- Calculate
- Classify
- Process
- Draft
- Review
- Validate
- Approve
- Reject
- Override
Your instruments already produce the data.We automate what happens after.
From images and instrument exports to QC, analysis, and reporting, Pelushka's Lab works with the data and workflows your laboratory already generates. We don't replace your instruments. We automate the repetitive work around them.
STAGE 05: DATA
// Ingestion & NormalizationThe work your lab already performs
Laboratory Network Attached Storage, Cloud Lakehouse, Object Stores
Raw measurement data, sensor logs, and plate manifests converge into shared infrastructure.
Where manual effort slows the workflow
Unstructured Schema Silos
Measurement tables, raw image folders, and plate manifests exist in disconnected directories with inconsistent coordinate systems and missing batch metadata.
Automation built around your workflow
Maps heterogeneous machine outputs into standardized columnar schemas, linking every raw pixel and intensity score to its originating sample ID and run conditions.
Works with your existing tools
Built around the data, files, and software your laboratory already uses, without requiring a wholesale system replacement.
Reproducible by design
Every step and transformation is tracked so results you generate today can be reproduced tomorrow.
Scientists stay in control
Automation detects, organizes, calculates, and drafts. Scientists review, approve, reject, or override any step.
From one bottleneck to laboratory infrastructure
Find a bottleneck, build the solution, validate it, and learn from it. Over time, successful automations become reusable infrastructure.
One repetitive task
We find a single bottleneck and build software to remove it.
One automation
The solution becomes part of the workflow and proves its value in practice.
Connected infrastructure
Multiple proven automations can, over time, become a connected laboratory data layer.
Use the simplest technology that solves the problem
We choose tools based on the bottleneck, not the other way around. AI is one option, not the default.
Request a Free Workflow Assessment
Tell us what your scientists currently do manually. We'll look for the parts that can be automated.