Evrone

Invariance is something we love. Sometimes you don't need to create a new product to simplify life and optimize the work of an entire team. That's exactly what we did at Evrone when we improved our time tracking workflow – we took an existing system and customized it to fit our needs. That's how Timesheet came to be.

We use Yandex.Tracker for tracking work hours. The system works well and is convenient, but there were nuances we didn't really like. For example, we wanted not just to log time, but to manage entries, edit them, and see the full picture without unnecessary effort. Wait for Tracker's creators to improve it? Not our style. We simply built our own tool inside it.

We built in everything needed for smooth work:

▪️ Detailed reports on tasks and projects broken down by day, week, and month

▪️ The ability to edit time entries

▪️ Automatic loading of active tasks

▪️ Filters by status, queues, and projects

▪️ Support for two languages and time zones

▪️ Input in minutes by default — it really is simpler

And yes, we know: time tracking issues in Tracker aren't something only our team faces. That's why Timesheet is open source – it's open to everyone. Any team can take the code, customize the logic or interface to fit their own workflows.

For us, this project is an example of how customizing existing tools helps build transparent development processes. Sometimes you don't need to reinvent the wheel at all – you can just tweak it a little.

You can learn more about the project here.

We published our BeRealFan project on Behance. Thought I'd share a few thoughts on how we approached it.

Designing complex digital products is really about working within constraints. With BeRealFan, the tricky part wasn't making it look good – it was building one coherent interface across a bunch of different systems: microservices, payments, streaming, messaging, analytics. Each of them has its own logic. Design is what ties it all together into something that actually feels seamless.

The visuals are based on patterns people already know from social platforms – clear hierarchy, familiar interactions, consistent navigation. But there's also a bit of personality: subtle details and a lighter tone that make it stand out.

End result – a product where design helps manage complexity, not add to it.

More visuals on Behance.

When it comes to fintech and crypto, many expect magic or overly complex systems. At Evrone, we take a simpler approach: if a task can be solved with classic tools without compromising quality – that's how we solve it.

The txn.pro project is a perfect example. It's a B2B service for sending and receiving crypto payments with automatic fiat conversion. The client needed a reliable, working product that wouldn't fall apart and could scale easily. So we built a minimum viable product with the right architecture from the start.

The backend is Ruby on Rails. Blockchain and liquidity integrations were moved into separate modules – keeping the core isolated and allowing load to scale independently. We chose a modular monolith: if the project grows, we can split it into microservices seamlessly. Our DevOps engineer planned the infrastructure from day one: DigitalOcean for staging, AWS for production.

The frontend is React with TypeScript, using a DDD approach – split into clear layers so development stays fast over time. For the design, we presented two concepts: one with dense, Excel-like data (familiar to experienced users) and a more creative alternative. The client chose the first – giving txn.pro its confident, industry-standard look.

Even in a dense interface, we kept beginners in mind. We added subtle hints and broke complex scenarios like 2FA setup into steps so users wouldn't get lost. Special attention went to transaction validation – in crypto, mistakes are costly, and the interface should prevent them at every turn.

The project is live, running smoothly, and we continue to provide support.

No uniqueness for the sake of uniqueness. Just a clear tech stack, clean architecture, and no headaches for the client during maintenance. If you need a crypto gateway or solid fintech development – we at Evrone deliver without surprises.

👉 More details.

Merch can't be great? At Evrone, we're proving otherwise – because we always create Internal merchandise items exactly the way our employees want them: comfortable, stylish, and practical.

This year's set includes a jacket, a thermos, and a travel mug. No flashy branding, no unnecessary details. The focus is on things that fit seamlessly into everyday life: office, commute, meetings, trips. The jacket is lightweight, windproof, water-repellent, and doesn't restrict movement. The thermos and mug simply keep drinks at the right temperature and feel comfortable in your hand.

At conferences, people constantly come up to us, ask where these items are from, and are surprised to learn it's our internal merch – and that it's not for sale.

To get great photos, we did a photoshoot with our designer who was involved in creating the merch. When it's a real team member wearing the items rather than an abstract model, it adds honesty and removes the distance between the object and the process.

The photos look like full-scale product shots – clean lighting, precise angles, attention to detail. But they're based on real items shot on a phone, followed by careful visual refinement using AI tools.

More photos of the merch we're proud of here.

Recruitment Automation: Killing the Manual Salary Transfer

Everyone talks about AI in HR, but getting it to actually work without weird mistakes is harder than it looks. We took one concrete problem and solved it.

The problem

Our ERP pulls data from Huntflow and job boards. Recruiters were losing 2–3 hours daily just moving salary numbers around. Open a resume, hunt for the digits, jump to another system, convert currency, standardize the format, paste it in. And if the data was off, finding the right person for a budget was pure guesswork.
Plus candidates write their expectations in a hundred different ways: "120+ net", "$5k-7k", "looking for 180+", "min 150", or just "open to your range". All had to be normalized by hand.

The fix

Template parsers died immediately – resumes are too messy. We used Qwen with structured output. Feed it a raw resume, get back clean fields: role category, timezone, English level, grade, location, salary.
Huntflow was worse. The text there is full of junk like "200/year" or "$3000". Regex was useless, base LLMs kept messing up. So we grabbed a dataset from old records, annotated it, and fine-tuned YandexGPT 5 Lite with LoRA – trained specifically to get amount and currency right.

The outcome

Huntflow lookups dropped 90%. Data accuracy hit 95%. Recruiters now manually check only 5% of cases – the ones with truly bizarre formatting. Time that used to vanish into copy-paste hell now goes into actual recruiting.

Want the same for your team? We're here to help. Full case here.

Our go-clean-template has crossed more than 7,000 stars and 600 forks on GitHub — and the numbers are still climbing.

It was built to solve a specific problem: how to start a new Go service and not end up with entangled, hard-to-maintain code a year later. The goal is a foundation that establishes a clean Clean Architecture structure from day one—isolating business logic and keeping the project scalable.

The community response shows this approach resonates. Behind every fork is a real project where the template helps maintain architectural clarity and development speed.

The project is actively maintained: over 40 releases, the latest being v1.14.1 (October 2025). This isn’t “set and forget”—it’s something we use ourselves and keep up to date.

Start your next project with a solid foundation → github.com/evrone/go-clean-template

How do you help a media giant not just renovate, but build a new home? We faced this exact challenge. This case is clear proof that a well-assembled team is what makes a project succeed.

VentureBeat approached us. It was a challenge. Working with a leader in the tech media industry means there is no room for error. Their platform was obsolete and needed replacement, but with one non-negotiable condition: 6 million readers could not lose access to 20 years of content. The task was complex: expand editorial capabilities, reduce costs, and find a solution for an archive of over 130,000 articles. And all of this – within the tight deadlines set by the client from the start.

We proposed a transition to a headless architecture using Contentful. This move created a modern and flexible foundation. Editors gained speed, the site got a new interface, and overall performance improved significantly.

The key challenge was migrating the archive. A full migration of all 130,000+ articles into the new system was economically impractical. We implemented a hybrid model: new, active content runs on Contentful for speed and flexibility, while the historical archive resides in optimized cloud storage. This decision reduced their ongoing CMS costs by over 20%.

In parallel with developing the new design, we built custom scripts for secure data migration. The result – a seamless transition with no downtime. Now, VentureBeat has not just an updated website, but a reliable and scalable platform ready for future challenges.

If you also see that your media platform no longer meets modern requirements – let's discuss potential solutions.

We solved a classic GPU underutilization problem for a European research lab. Their high-performance hardware was largely sitting idle. The root cause? Their system reserved an entire graphics card for every single task, large or small. Most workloads ended up using only 10-20% of the available power. Multiple teams then competed for these underutilized cards through a slow, manual management process. The result was zero visibility and zero efficiency.

Our approach? We didn't guess. We tested. On their actual, heterogeneous GPU fleet, we implemented and evaluated two strategies: partitioning the resources of a single GPU and intelligent time-scheduling. The data from these real-world tests gave us the correct configuration.

In eight weeks, we delivered. We built them a custom MLOps platform on Kubernetes, utilizing open-source tools like Ansible, Grafana, and Ray. This transformed their static hardware into a dynamic, shared resource pool. Now, tasks run in parallel, utilization is high, and they have complete control. There is no vendor lock-in. They own the platform outright.

The lesson is straightforward: often, you don't need more hardware. You just need to use properly what you already have.