Services/R&D
Research and Development
When nobody can say in advance whether an idea will work, we build a prototype and find out, before you commit to the full product.
Questions we help answer
Seven questions clients bring, in any industry. Each of them can be answered with a small test, before the budget of the full product is spent.
Can it be built at all?
It may be a new idea, an algorithm, a connection to a device or an unusual interface, so we build the risky part first and alone.
- Yes, yes with limits, or no, and the reasons
- What the full product would take
- Which part of the test the product can build on
Can it be automated?
A process is done by hand today, such as sorting requests, checking documents or entering data. We test whether software can do it, and how well.
- Which steps can be automated and which still need a person
- How often the automatic result is right
- What it saves in time and money
Which technology should we choose?
There are two or three ways to build the same thing, so we compare them on your task, not on general opinion.
- Speed, cost and risks of each option, side by side
- How each one behaves on your data
- A recommendation with the reasons
Will it work with our systems?
The new product has to exchange data with the systems you already use, so we test the connection before the product depends on it.
- Whether the system gives access to the data you need
- How fast and how reliable the exchange is
- What has to change on either side
Is our data good enough?
A product that learns from data or reports on it is only as good as that data, so we check what you have before anything is built on it.
- Whether there is enough data, and whether it is clean enough
- What has to be collected or put in order
- How long the preparation takes
What will it cost to run?
A new technology looks cheap in a demo, but we count what it costs at the volume you expect.
- The cost of one operation and of one month
- Where the cost grows faster than the audience
- Cheaper ways to get the same result
Build it or buy it?
A ready-made service may already do a part of the job, so we find out which part.
- What the ready-made solutions cover
- Where they fall short for your task
- The cost of each path over time
Four ways to run the research
The format depends on how much is unknown, so we propose one after the first call, together with the time and the price, and you approve it before we start.
Feasibility study
This is the shortest format, because it answers one question: can this be done with the technology and the data you have? We study the task, try the key part on a small sample and then write the conclusion. If the answer is no, the conclusion also names the conditions under which it would change. As a rule, it takes one to two weeks.Proof of concept
A proof of concept tests the risky part in practice, before the product is built around it. We build the smallest working thing that can answer the question, and then measure it on your real data: quality, speed and cost. As a result, you get numbers instead of a promise, and also the code of the experiment. As a rule, it takes two to six weeks.Prototype
A prototype is a working model, so you can show it to users, partners or investors. It includes the risky part and also enough of the interface around it to try the idea the way people will use it. Because it is built to learn quickly, a part of it is rewritten later, and the conclusion says which part. As a rule, it takes one to three months.Technology comparison
Sometimes the question is not whether to build, but on what. We take two or three options, such as different frameworks, libraries or cloud services, and run a small test of each on your task. Then we put speed, cost and risks side by side. You then get a recommendation with the reasons, usually within one to two weeks.How a research project runs
Six steps that keep research from turning into an endless project.
Not sure whether your idea can be built?Tell us the question. We will then propose a way to test it and name the cost of the test.
Discuss a prototypeProjects where the answer was not known in advance. Each case shows the stack, the task and what was hard.
All case studies
Photo recognitionTruck parts identification appPhoto recognition on the phone: the app names heavy-truck parts from a picture, on iOS and Android.Computer vision · Flutter · Python
New groundWaxify: NFTs with physical goodsA public Shopify app that sells NFTs on WAX together with physical goods.Shopify · Node.js
AlgorithmsTezos trading botA C++ rewrite of a trading bot: pool monitoring, a price feed and a rules framework.C++
Engagement models
Three ways to work with us. We help you pick one on the first call.
Our engineers work as an extension of your team, under your lead or ours.
We agree on the scope, the price and the deadline before the work starts.
You pay for the hours worked and change priorities whenever you need.
We turned to the 24Gen team when the project was in a critical stage. Deadlines were looming, and the product was far from ready. From the very beginning, we were impressed by the team's professionalism and accountability. In a short time, they assembled an MVP, set up scrum processes, and introduced an AI-based solution for moderation. Despite tight deadlines, all key modules were delivered on time and with solid quality. We are very satisfied with the collaboration and confidently recommend 24Gen as a reliable partner for large-scale projects.
Watch the video
From our blog
All articlesWhat if the answer is that it does not work?
Then you have saved the budget of the full product. You also get the conclusion and the code of the experiment, so you know why it did not work.
How long does a research project take?
A feasibility study or a technology comparison usually takes one to two weeks, a proof of concept two to six weeks, and a prototype one to three months. We set the time and the budget for each question before we start, and you approve them. When they run out, you get the answer we have.
Can the prototype become the product?
In part, because a prototype is built to answer a question quickly, so some of it has to be rewritten. The conclusion says which part.
Do you work with our data?
Yes, and under an NDA when you need one. Therefore the data is used only for your project.
Do you build the product after the research?
Yes. The same team continues, so nothing has to be explained twice.
How do you estimate the cost?
After the first call we break the task into parts and send you a written estimate. When the scope is clear, we fix the price and the deadline. When it is open or likely to change, we work on time and materials or as a dedicated team.
Who owns the code? Do you sign an NDA?
You do. The code and all rights to it pass to you in full. We sign an NDA whenever you need one.
Tell us about your product
Describe the task in a few lines. We will get back to you to set up a call.
info@24gen.space



