Artur Pan and Madiyar Bakbergenov had both built successful careers in the IT industry when, in 2024, they decided to launch PanDev Metrics. The product analyses how IT teams work, calculates how much money is spent developing individual features, and identifies signs of employee burnout. The young company has already raised $400,000 at a $5 million valuation and is now preparing for its next funding round.
As part of “100 Startup Stories from Central Eurasia”, a joint project by Digital Business and Astana Hub, Madiyar and Artur shared how they invested the $100,000 they had originally saved for a flat into the startup, and why some developers may not be too happy about the service. We also talked about why the founders are already considering an exit and what they see as the next stage for PanDev Metrics.
“I had a choice: buy a flat or invest the money in my own project”
– What did you do before PanDev Metrics, and what made you decide to launch a startup?
Madiyar: – I’m from Karaganda. I joined Chocofood in 2019, and a year later my wife and I started our own business, a women’s clothing shop. In 2021, I was invited to move to Almaty to help grow Choco’s restaurant business. After that, I joined Kwaaka, a startup that connects food delivery platforms with restaurant POS systems. We made strong progress in just one year, growing from 200 food outlets to 1,200 across three countries. I then worked on a project in Indonesia with 160,000 monthly active users. Over time, I moved into consulting, helping startups develop their go-to-market strategies, improve their decks, and sharpen their pitches.
Artur: – I’m from Almaty. After college, I skipped university and went straight into work as a developer. Over time, I became one of the first ten engineers who helped build Kaspi Store and the courier app. The feature where you give the courier the code from your text message to confirm delivery was one of the things I worked on. After Kaspi, I joined Vlife as CTO, working on a loyalty programme for petrol stations. It now has 1.5 million users. I then moved to Umico, now Birmarket, an Azerbaijani digital platform that brings together a marketplace, a loyalty programme, and fintech services. I was part of one of its largest teams, with nearly 150 people in the department. The development process there was not very transparent. That was when I came up with the idea of building an analytics platform that could show how development was progressing, how much time different tasks were taking, and how each specialist was contributing to the project. I wanted to avoid situations where 20 people were doing work that could realistically be handled by three.
Artur Pan
I put $100,000 of my own money into the project, which was everything I had saved for a home. I had a choice: buy a flat or invest in my own startup. I chose the startup. I put a team together and started building PanDev Metrics. The money was running out fast, but then we made our first sale for $55. At the time, I did not even know how to accept the payment. Our first customer was a small outsourcing company from Kazakhstan.
When things became really difficult financially, I connected with Madiyar on LinkedIn. I already had the product, while he had experience in sales and growing startups. We quickly realised that we could make the project stronger by working together, and Madiyar joined PanDev Metrics as a co-founder.
Madiyar: – Artur and I first met in person at CEVF 2025, a major venture forum for startups and investors. That was where they announced which companies had been selected for the accelerator and would receive $200,000 in funding from MOST Ventures. The wait was nerve-racking, and PanDev Metrics was the very last name they called. After that, we joined the accelerator and sorted out all the organisational details. We set up a legal entity and opened bank accounts, registered with the Astana International Financial Centre, and became an Astana Hub member.
Madiyar Bakbergenov
“None of our clients uses PanDev Metrics as a stick to beat their teams with”
–What is the main idea behind PanDev Metrics?
Artur: – The IT industry is a classic case of the cobbler’s children having no shoes. A company might have digitised everything, from its call centre and sales to content management, while the way its own developers work is still far from transparent.
A lot of business owners see developers as wizards who do something highly technical that nobody else really understands, so they tend not to interfere. But how many developers do you actually need, 100 or 50? Who is genuinely putting in the work, and who is only working a couple of hours a day? Our system gives you the answers.
We give companies a clear picture of how their developers are working, how long different tasks take, and how much each feature costs to build. PanDev Metrics is one of the few tools that can put a real figure on development costs and show where an IT department’s budget is going. You enter each developer’s salary into the system, and it works out the cost of individual projects, features, and code branches. We have also started building a recommendations feature. The platform will analyse the data, spot weak points in how IT teams are set up, and suggest where things could be improved. We plan to roll this out soon.
We can see how quickly each engineer works and use that data to estimate how many person-hours a similar task would take. For example, around 40,000 hours have gone into building PanDev Metrics so far, including 25,700 hours of actual coding. Many companies cannot even give you figures as basic as these.
– What exactly is PanDev Metrics? Is it a piece of software that integrates with a company’s existing ecosystem?
Artur: – Engineers already work with tools such as Claude Code, along with dozens of other platforms. PanDev Metrics connects to those tools as a plugin and analyses how they are being used.
– It sounds a bit like keeping tabs on developers. Have you faced any negative reactions?
Artur: – Quite a lot, actually. The platform makes developers’ work much more visible, and some people are not too happy about that. There are developers who juggle several jobs at once, and of course they do not want anyone finding out. It can also catch out people who use Claude Code or ChatGPT to write code and then pass it off as their own.
That said, none of our clients uses PanDev Metrics as a stick to beat developers with. For one thing, the system is actually quite developer-friendly. It does not rush engineers or push them to work harder. It simply looks at the facts. It also gives product managers and team leads a clearer picture of who is doing what, so they can create a fairer and more transparent system of rewards and incentives. We have all heard stories like this: one quiet person carries the whole project, but nobody notices, while someone else talks themselves up and ends up in management’s good books. Employees know the system is being used. This is not secret surveillance.
– Can you think of a time when PanDev Metrics helped uncover a problem in a business?
Artur: – At one company, we discovered that engineers were still working on a project that had already been shut down. The developers simply had no idea that the company had dropped it. We have also seen cases where companies were spending far too much money building products from scratch when an off-the-shelf solution would have been a much better option.
In another case, management could see that the developers were not getting much done. When they looked into it, they found that the project manager was not describing tasks clearly enough, so the developers were spending a lot of time trying to work out what was actually expected of them. As soon as the manager was replaced, productivity went up.
Interestingly, our system often shows that the developers are not the problem at all. It is the way the work is organised.
“We’d love it if PanDev crashed under the weight of all that traffic”
– What makes your product different from the others out there?
Artur: – The biggest difference is accuracy. Most CTOs only have data from Jira and Git to go on, so they are looking at task tracking and code changes. PanDev Metrics goes much further. It connects to the full range of tools developers use every day, including their IDE, browser, terminal, and AI tools.
We are also adding AI features. For example, the system is learning to spot possible signs of burnout from the way an employee writes prompts. Say someone starts speaking to the AI in a rude or negative way. That could suggest they are under stress. Another warning sign is asking the same questions over and over, which may point to poor concentration. Even neutral responses such as “it didn’t work” or “it wouldn’t start” are taken into account. The system also looks at whether someone is working late at night or coding far more than everyone else. There are many different signs that could indicate someone is overloaded.
– Does the platform work with all the tools developers use?
Artur: – Pretty much. Even when a tool is not supported yet, we can usually add it within a week if a customer asks for it.
– Is there a limit to the size of company PanDev Metrics can work with? For example, would it be just as effective in a company with 10,000 employees as it is in a small startup?
Artur: – We have not yet rolled the product out at a company with 10,000 engineers, but we have run load tests. The system handled the data flow from 5,000 virtual employees without any problems. To be honest, we would almost be happy if PanDev Metrics crashed under the load. It would mean we had landed a client that large.
– PanDev Metrics analyses how developers work. Do you use AI for that?
Artur: – There is actually very little AI at the heart of the product. We are starting to add more AI features, of course, but the core function, measuring how developers are performing, does not rely on AI.
A lot of companies are using AI extensively now, but very few can tell where their tokens are actually going, whether they are being spent on work or, say, looking up recipes.
Sometimes the figures are completely mind-blowing. According to unofficial reports, Meta used 73 trillion tokens in a single month, which works out at around $220 million. And that is before you even consider the environmental impact, including how much electricity and water went into running and cooling the data centres.
A lot of this can be done without AI, which is important for our clients because it makes the platform much easier to integrate. Not every company has the infrastructure or powerful GPUs needed to run AI, and some simply cannot send their data to third-party providers. From our side, we only see how many users there are and whether the subscription is active. That makes it much easier for the platform to get through a company’s security review.
“We’re almost breaking even”
– How many people are on the team?
Artur: – At the start, I mostly hired junior developers, and AI was nowhere near as advanced as it is now, so the team was actually bigger back then. We started with 12 or 13 people, many of whom I mentored and trained myself. Now that AI models have become much more capable, and with our focus on efficiency, we work with six experienced engineers.
We also have a business development manager, a lawyer, an accountant, and a finance specialist.
– Last autumn, you raised $400,000 at a $5 million valuation. Are you still growing on that funding, or is the business already profitable?
Madiyar: – We are still running on the funding we raised, but we are very close to breaking even. We expect to get there within the next few months. A few large clients are now moving from pilot projects to paid plans, and our current MRR is $20,000.
We are also gearing up for another funding round and plan to raise between $1.5 million and $2 million. The goal is to expand into the US and Europe.
– Who are your clients, and how many do you have?
Madiyar: – Within just a couple of weeks of raising the investment and registering the company, we signed 15 to 20 clients, all through personal connections. Around 85% are based in Kazakhstan, including MyCar, Biometric, Halyk Finance, MyLink, Damumed, Zeely.ai, ABR Tech, Freedom Delivery, Freedom Restaurants, and others. The rest are from Uzbekistan, Kyrgyzstan, Azerbaijan, and Tajikistan. We also have one client in the US.
Our pricing is $500 for up to 20 developers, $1,000 for up to 50, and $2,000 for up to 100. For teams of more than 100 engineers, we agree the price directly with the client. Most customers go for the package covering up to 50 employees.
We have big plans for growth. We are working with distributors who will help promote the product, and we are mainly targeting the US and Europe, especially Canada and the UK, where we have already made a lot of useful contacts. We also reach out to companies by email and on LinkedIn, using AI agents to make the first contact and set up meetings. Around 9% of that outreach turns into a meeting, which is a pretty good result for this kind of channel. Personal connections in the IT industry help too.
– As well as driving sales, you are building partnerships with major industry players. What does your partnership with JetBrains bring to the business?
Madiyar: – We are building the startup with the aim of eventually selling it to a major player in the market. That is why it is important for us to stay close to companies such as JetBrains and Atlassian.
PanDev Metrics has joined the JetBrains Consulting Partner Program. This means we can help companies roll out JetBrains tools and use data from their development environment for analytics, giving them a clearer view of how coding work is progressing, where teams are losing time, and what could be improved in their processes.
– Why are you planning to sell the startup?
Madiyar: – A lot of successful products in our space have eventually been bought by major tech companies and folded into their wider ecosystems. Atlassian’s acquisition of Trello in 2017 is a good example. For a product like ours, that is a pretty natural route. It could fit well into a larger company’s product range and add real value to what they already offer. It also makes sense from an investor’s point of view, because they usually back startups with the hope of an eventual exit and a return on their investment.
– When do you think you could be ready for a deal?
Madiyar: – We are currently in talks with a strategic player and discussing which metrics would matter most in a potential deal. We cannot share any details just yet. We are looking at a two-to-three-year timeframe. By then, we expect to have reached the numbers they are looking for.
– Let’s imagine we’re having this conversation again in a year. What do you think the company will be worth by then, and what new developments will you be able to share about the product?
Madiyar: – I think the company will be valued at between $25 million and $50 million a year from now.
Artur: – We hope to have signed several hundred contracts by then and for PanDev Metrics to be seen as the leading product in its space, at least in Kazakhstan. We want it to become one of those tools every CTO relies on. Jira and GitLab are industry giants used by almost every company. Our ambition is to become just as widely adopted.