We’re sitting down with Bohdan Snisar, an accomplished CTO and engineering leader with 13 years of experience in the software industry. Originally from Ukraine and now based in London, Bohdan has built an impressive career spanning from software engineer to CTO roles at companies like Revolut and Wix. With a B.S. in Computer Science and M.S. in Economics, he’s founded multiple startups and currently serves as CTO of a stealth startup building a self-updating documentation platform. His expertise spans GenAI systems, financial technology, MLOps, and large-scale infrastructure, making him a fascinating voice in today’s tech landscape.
1. What initially drew you to software engineering, and how has your perspective on technology evolved over your 13-year career?
I was drawn to software engineering initially by the blend of problem-solving and the ability to build systems that scale and impact millions of users. My own network started at SoftServe in 2012, when I was intrigued by the manner in which code is the solution to building real business value. My own vision has grown over the years. I have transitioned from a technologist’s view of technology to how technology influences overall business objectives and user requirements. My M.S. in Economics on top of the Computer Science grounding further expanded my vision, now which makes me consider technology not merely an engineering problem, but a way to create scalable solutions with actual business value. Now, whether building trading infrastructure for Revolut or GenAI frameworks, I address problems through a multi-disciplinary approach that includes both engineering and strategic business needs.
2. You’ve transitioned from individual contributor to CTO across multiple companies. What were the key leadership lessons you learned during this progression?
The biggest lesson I’ve learned is that leadership isn’t about being the best engineer, it’s about enabling others to do their best work.
I came to understand this while at SoftServe, and even more so at Wix and Revolut, where I had good teams under high stress. I came to understand that my role was not necessarily about my own contribution but rather about setting direction, unblocking teams, and how to step in and when to get out of the way.
Now, as a CTO, I still write code, mostly Python, TypeScript, and Go, but it’s mainly about getting the team aligned, setting technical direction, and making sure we’re making the right thing, in the right way.
3. Your current stealth startup focuses on self-updating documentation. What problem are you solving?
The problem we’re solving is outdated, disconnected documentation, a major source of technical debt and developer frustration.
In most teams, code moves faster than docs. This leads to confusion, bugs, and wasted time trying to reverse-engineer intent from the codebase. Our dev tool keeps documentation in sync by automatically regenerating it from source code using a language-agnostic parsing layer (Tree-Sitter + Stack Graphs) and LLMs like Mistral and Llama.
Instead of documenting as an afterthought, we make it a living part of the codebase – always up to date, always applicable. This reduces human error, accelerates onboarding, and keeps big teams aligned without the extra documentation burden.
4. You built trading infrastructure in Revolut, you worked at a big scale at Wix. What were the key architectural or engineering decisions that enabled the system to scale reliably?
Compliance, transparency, and reliability were the focus areas in Revolut. We designed clean, auditable portfolio management and trading services. They were all fault-tolerant and easy to trace back – a requirement in a world that is regulated. There were risk controls added to maintain control over the system even when pushed to stress.
At Wix, scale was the priority. I contributed to a search system handling millions of requests per minute on 350M+ documents. We achieved p95 latency under 100ms via smart caching, parallel indexing, and optimized query paths – all tunable for scaling without sacrificing performance.
5. You’ve worked extensively with GenAI and LLMs across multiple ventures. What’s your take on the current state of AI in enterprise applications versus the hype?
The gap between GenAI hype and real enterprise adoption is wide, and I’ve seen it firsthand while building multiple AI startups.
At Redress, we created a GenAI fashion stylist that was 44% better than Farfetch’s benchmark on blind user tests. But only because we limited the scope, stuck to structured outputs, and tuned the system to actual user needs. At 0docs, we are addressing the stale docs problem by re-generating docs from source code using Tree-Sitter and LLMs like Mistral, but even there, it’s not plug-and-play.
Most companies want GenAI to be a magic switch. But the reality is: it only works when well integrated into workflows, well-constrained, well-secured, and with complete understanding of where LLMs can (and cannot) generate value.
6. Your experience spans from traditional software engineering to cutting-edge AI systems. How do you stay current with rapidly evolving technologies while maintaining deep expertise?
I stay current by staying hands-on by constantly testing out the newest implementations and using them on real-world problems.
In my startups, I don’t merely keep pace with trends: I prototype, I benchmark, I integrate. From leveraging Tree-Sitter for language-agnostic code parsing to optimizing Mistral for self-updating docs, I learn through building.
Being in incubators like NVIDIA Inception and Google for Startups gives early access to new technologies, but the secret is to filter out what’s appropriate and go deep when it maps to the product. I stay alert by combining real-use scenarios, technology interest, and collaboration with high-performance teams.
7. You’ve founded and co-founded several startups while also working at established companies like Wix and Revolut. What drives you toward entrepreneurship, and how do you balance innovation with execution?
My driving force in entrepreneurship is being able to build from first principles, free from legacy, and solve problems head-on at the root.
In startups, I can move fast, test assumptions quickly, and architect systems just for the problem in hand. That autonomy is very motivating. And yet, my experience at Revolut and Wix showed me the elegance of operating discipline – scalability, risk management, and clear processes.
And so I bring both: the speed and innovativeness of a startup founder, and the habits of execution of an engineer who’s shipped at scale. Innovation only succeeds when paired with rigor – sound testing, monitoring, and deployment practices from day one.
8. In your role at Wix, you achieved a 7% improvement in key product KPIs. What specific engineering or product decisions drove that result, and was AI involved in any way?
The increase in KPIs resulted from building a hybrid text search solution for Wix Ads Funnel. By using BERT embeddings for semantic intent along with the usual TF-IDF text search for literal keyword search, we were successful in creating a more effective search that could understand the nuance of user searches. Artificial Intelligence entered the scene a lot here because BERT embeddings gave semantic accuracy to the search, but the real genius was in bringing these AI algorithms together with the old-school ways of searching in an effort to gain speed and relevance. This hand-in-hand blend created more user engagement and higher ad targeting, and this impacted KPIs directly on big products.
9. You’ve worked with teams across different countries and cultures, from Ukraine, the US, EU to London over a long period. How has the software engineering role transformed in the current days?
The biggest revolution for the software engineering industry is embracing remote-first and distributed teams. Remote working was not traditional when I started in 2012, but now it’s the norm. Engineers today need to be highly collaborative and super effective working remotely, and this has transformed the collaboration tools and culture we use. Furthermore, the work itself is multidisciplinary today.
Engineers today need to know a little bit of everything from coding to DevOps, security, and AI. The technical complexity of the systems has increased too, and the engineers are required to work in multiple domains simultaneously, either creating a RAG architecture in Wix or releasing language models within my new company.
10. Looking ahead, what emerging technologies or trends do you believe will fundamentally change how we build software in the next five years?
Agentic systems, LLM-native interfaces, and code-aware models will shift software development from human creation to collaborative orchestration. Programmers will shift from boilerplate coding to higher-level system guidance, with AI taking more of the heavy lifting like code generation, test scaffolding, even architectural suggestions.
In addition to that, I think autonomous infrastructure and real-time feedback loops are enablers. Declarative DevOps with self-healing infra, CI/CD with LLMs integrated, and monitoring and adjusting to live system behavior tools will fundamentally alter the way we build and ship.