There’s a trade-off that comes with seniority in engineering that nobody really advertises: the more you lead, the less you build.
For Wrike Engineering Director Aleksei Kartavenko, his leadership role meant less hands-on work — until AI changed that.
With over 25 years in the industry and almost a decade at Wrike, Aleksei leads two product units, and his passion remains solving customer problems and building great products, no matter his title.
“For a time, the heavy managerial demands of a director’s life meant I lost the opportunity to contribute personally to this work. Recently, AI has given me back that personal touch. I now regularly have tools like Codex or Cursor open, allowing me to do work ‘by hand’ again. I can kick off a prompt, attend a meeting, return, correct the result, and essentially create new things in the background without impacting my primary responsibilities.”
That idea, “Create new things in the background,” captures Wrike’s current AI culture.
A company already in motion
Wrike was ready when generative AI went mainstream. What’s changed recently is AI’s broader reach inside the company.
“Our company has long had a dedicated Machine Learning (ML) department, predating the recent surge in generative AI. However, this expertise was often isolated, and other R&D departments rarely engaged with it. For over a year now, we have been in a total AI transformation mode, a movement I’m part of as an Engineering Director for two product units. We’ve made powerful AI tools available to every single Wrike employee.”
“Furthermore, we’ve developed and actively maintain comprehensive internal resources tailored to our business needs, expertise, customer feedback, and codebase. We actively promote and encourage the use of AI every day. Personally, I have access to all the necessary tools and can request more for any experimental needs.”
For Aleksei, that access is tangible and daily. His work runs across a full stack of internal and external tools.
“My daily work heavily relies on a variety of internal and external AI tools. Internally, I use our AI portal, which provides access to knowledge bases, and leverage Wrike’s AI features, including agents, automations, summarization, and prioritization tools. For brainstorming and knowledge retrieval, I turn to external chat models like ChatGPT and Gemini. When it comes to the actual coding process, I use specialized applications such as Cursor and Codex.”
But infrastructure isn’t the whole story; culture matters even more. Let’s delve into what makes this environment different.
What actually makes experimentation feel safe
Some companies say they encourage experimentation. Fewer actually make it frictionless. Aleksei has been at Wrike long enough to know the difference.
“I’ve been with Wrike for nearly a decade — long before AI became a widespread topic. What has always fascinated me is how easy it is to launch a new initiative or experiment with a new idea here. I’ve never experienced pushback from my boss or stakeholders; every meaningful proposal has always been greenlit. That supportive environment is what makes working here so interesting and rewarding.
“Regarding AI, I have the necessary tools and do not need to create detailed expense reports for specific projects or experiments. Instead, any work that yields an impactful outcome is celebrated and fully backed.”
That last point matters more than it might seem. The absence of bureaucratic friction, no lengthy approval chains, and no project-by-project expense justification creates the conditions for genuine experimentation. You can try things. You can move fast. And when something works, it gets recognized.
This culture of experimentation leads to real results, including one standout week that began with a migration, cost just $30, and produced a lasting impact.
The $30 win
Wrike has an internal tool called Dossier used daily across the company for people and org information. It was in need of an upgrade, so Aleksei took it on as an AI-powered experiment.
“Using AI, I successfully migrated our popular internal tool, Dossier, from an outdated stack to a modern one in one week. The process involved careful prompting based on an AI-generated plan, costing about $30 in token credits. This rewrite now allows me to fulfill new feature requests in one to two hours, enabling daily releases without impacting my capacity. This experience is driving my ideas for applying this cutting-edge approach to our main product work.”
Not bad for a $30 investment.
The harder problem nobody talks about
Still, Aleksei knows these kinds of wins are harder to scale in a bigger organization.
“Applying AI-driven efficiency to a large, complex product with 20 years of history, as I’ve experienced, is inherently more challenging than working with smaller-scale pet projects or internal applications. True transformation requires a careful, long-term process — analyzing and reworking every component of established enterprise-grade processes, or even challenging the business’s foundational pillars. This involves changing mindsets, optimizing collaboration loops, and identifying exactly where and why AI fails to deliver value.”
This is where experience becomes the actual differentiator. AI is only as good as the human directing it, and in enterprise development, that human needs to carry years of context, judgment, and domain knowledge that no model currently has.
So, where does Aleksei see his role growing?
“The speed of change is incredibly fast, demanding to be on alert all the time. I have a persistent feeling that my AI prompting and use of AI purely as an assistant are becoming somewhat outdated. Because of this, I’m keen to explore agentic workflows and orchestrations. This focus directly aligns with the direction of thinking within Wrike R&D and the company as a whole.”
Tips for beginners in engineering
Aleksei has some advice for those early in their engineering career:
For devs: Immediately acquire and start using the new tools. Remember that while AI is an effective assistant, you must maintain control and the highest level of expertise. Enterprise development requires precision, not just ‘vibe coding,’ and your irreplaceable knowledge is key to building an effective partnership with your machine assistant.
For managers (especially with a software development background): Install and regularly use the tools. This hands-on engagement is essential for understanding the technology’s current capabilities. You might discover that direct contribution to the product is more accessible than you currently perceive.
Staying in control is key: these tools should enhance your expertise — they can’t replace it.
Curiosity off the clock
At Wrike, personal initiative and company direction go hand in hand. Aleksei’s after-hours projects show his thinking in action.
“I actively pursue my passion for building products. A recent example is the successful launch of Rimata, my free modern Greek verbs app on both iOS and Android. I’m confident it’s the best on the market. Developing and launching this app would have been impossible without AI assistance, which helped with everything from coding and navigating store submissions to collecting and processing the vast database of modern Greek verbs. This app is currently used by over 3,000 consistent users and maintains a perfect 5.0 rating. This experience proves that high-quality products can be created by a single person. And then, while enjoying your success, you need to bring this experience back to the workplace and attempt to scale.”
For Aleksei, it’s clear that curiosity doesn’t clock out at 5 p.m. — building solutions is a constant drive. Wrike is where curiosity gets funded, experiments get approved, and engineers get to keep building.
He added one more thing, almost as an aside:
“And, yes, I stopped googling.”
In 2026, that might be the most telling line of all.
