Contact

Propose a project.

Kodulabor accepts a limited number of projects each quarter. We look for problems that are specific, measurable, and worth studying — where the experiment itself advances what we know about applied AI.

If the project fits our research agenda, we run the experiment, measure everything, and publish the results. You get a working solution and a structured assessment. The field gets another data point.

How projects are selected

Not every inquiry becomes a project. We evaluate each proposal against our research criteria: Is the workflow measurable? Is the problem real and recurring? Will the results be instructive for others? Projects that meet these criteria are scoped tightly — typically days to weeks, not months. Some are paid engagements. Some are research collaborations where the published case study is the primary output.

What we look for

—A specific, recurring workflow where AI might produce measurable improvement

—A content or knowledge-work bottleneck that follows recognizable patterns

—A build-or-automate question where speed-to-production is the variable under test

—A feasibility question — does AI actually work here? — that can be answered with real data

—A workflow where the traditional benchmark is known, making acceleration measurable

—Willingness to let us document and publish findings (anonymized if needed)

The common thread: the problem is real, the outcome is measurable, and the results are worth publishing.

How it starts

Describe the problem. We do a brief evaluation to understand the workflow and whether it meets our research criteria. If it does, we propose a scope, timeline, and assessment plan. If it doesn't, we'll say so — no cost, no pressure. We only take on work where the measurement will be worth publishing.

sergei[at]kodulabor.ai

Sergei Anikin

Selection criteria

✓A specific, repeatable workflow — not a vague AI strategy question.

✓Clear success metrics — time, cost, quality, or throughput that can be measured before and after.

✓Willingness to have findings documented and published (anonymized if confidentiality requires it).

✓Bounded scope — days to weeks of focused work, not open-ended engagements.