Engineering
Reports.
Strategy is easy on a slide. Here is what it looks like when it hits the friction of the real world.
Engineering Confidence at Scale
Business Unit Strategy LeadThe Context
Corsair is the elephant in the room of the PC component industry. It's known, familiar, and accessible in every serious computer shop in the world. Its components are a delight to DIY PC builders and performance enthusiasts. Me included, I've been building my own PCs since the 'beige box' era.
But for many customers, building a PC is an exercise in anxiety. Even an enthusiast might have built their last PC 3 years ago. 'Will this fit?' 'Is this compatible?' 'Do I have enough power?' What chance do new builders have? They often rely on a 'Sherpa', that one friend or cousin who knows PCs, to guide them through the process and validate their choices. We wanted to digitize that Sherpa.
The Friction
We had previously developed tools like the Custom Cooling Configurator, which were a guiding idea and proved that we can manage the technical complexity. We needed a mass-market tool that could guide a novice from an empty desk to a full $3,000 cart, managing compatibility for thousands of SKUs (including competitors') without overwhelming them.
The Architecture
We built the PC Builder, a logic-driven wizard that acts as a digital sales assistant.
The 'Sherpa' Logic: We worked with an agency to gut and rebuild a back-end infrastructure that was never meant to support this. We taught the system to understand relationships, not just specs. It calculates physical clearance, power draw, and cooling deltas in real-time. On every step, it filters out the noise and recommends the best match, plus a budget and premium alternative.
The Ecosystem Play: We included products Corsair didn't sell (CPUs and Motherboards), as well as competitor products like cases, because we understood that to sell the Corsair ecosystem, we first had to support the user's core choices. We were there to support a new builder, not just push SKUs. The 'Buy' button only works if the advice is honest.
The 'Artificial Friction' Insight
During internal testing, we noticed our dev partner was too competent. They made the tool too fast.
We were helping users calculate complex thermal and physical compatibility, and it was happening instantly. The result? Users wouldn't trust it. It felt lightweight, like a fake marketing tool that just pretended to do the work.
So, we slowed it down.
We added artificial loading screens that didn't do anything other than add friction. They told the user what we were calculating: 'Checking Thermal Load... Verifying GPU Clearance...'
Result: Trust skyrocketed. Conversion went through the roof. Sometimes, in consumer psychology, efficiency looks like latency.
The 90-Day Monolith
Head of Digital TechnologyThe Context
Nomago was born almost overnight as a market monolith—a merger of the country's largest public transport operators, a shuttle service, and a highly respected travel agency.
We suddenly had 800 rolling billboards on the road and a new brand nobody had heard of, but behind the scenes, it was a fragmented digital disaster. Oh, and there was a Croatian travel entity thrown in as well.
We needed to digitally merge these companies over the summer with zero downtime to public services. Established agencies wished us luck but refused to quote. The risk of a loud, embarrassing public failure (transit collapse, PR debacle) was simply too high.
The Friction
We had 3 months (90 days) to merge 5 companies with different backends: rigid government-connected transit systems, travel booking engines, and airfare search modules.
The biggest risk was SEO: The travel arm held the #1 spot for high-value keywords like 'airplane tickets.' Migrating them to a newly registered 'bus domain' risked wiping out years of organic authority and millions in revenue.
The Architecture
We assembled a lean team and built a 'Frankenstein' unified layer.
Headless Aggregation: We used a mix of APIs to ingest data from the rigid government infrastructure, travel booking engines, and a legacy WordPress library of travel packages, presenting a unified modern frontend to the user.
The SEO Fortress: We cataloged every single URL from the legacy brands going back years. We built massive, complex redirect maps in
.htaccessto tell Google: 'This isn't a dead link, it's a relocation.' It was a brute-force effort to ensure 1-to-1 content mapping.
The 'Suicide Mission' Reality
This wasn't a project about 'pretty design.' It was about survival.
We launched with zero downtime. Our dev instance was our live instance. We had 200k students hitting the site on Day 1 of the school year, and they had to be on time.
The Result: Organic traffic dipped for one month (standard for migration), then recovered to pre-launch positions for all target keywords. We proved that a scrappy, focused internal team can out-execute a large agency when the stakes are real.
Codifying the Senior Marketer
Co-Founder / Product ArchitectThe Context
AI promises a revolution, but the 'Chat' interface is failing businesses. Small business owners use ChatGPT to write 'marketing,' and they're just getting yelled at by their audiences who are frustrated with generic, hallucinated 'slop.'
The problem isn't the model; it's the user. If a user isn't competent in a field, they aren't qualified to instruct the AI.
Think of it like photography: I can ask Midjourney for an image, but because I don't have a photographer's eye or vocabulary, I get a generic result. The same happens when a non-marketer asks AI to write an email. They don't know the difference between a 'Launch Email' and a 'Nurture Drip,' so they can't ask for it.
The Friction
We needed to build a tool that bridges the gap between 'Admin Operator' and 'Senior Strategist.'
A tool that doesn't just 'generate text,' but offers the support a senior coworker would—asking the right questions and adding domain expertise to every communication.
The Architecture
We built SimplerWork as a Strategic Interface, not a chatbot.
Showing users what they can do: Instead of an empty prompt box, the system presents a menu of expert-defined outcomes (e.g., 'Onboarding 3-Email Drip'). This educates the user on what they should be doing.
The Expert Interview: The wizard acts like a human strategist, asking specific questions to gather context (Target Audience, Pain Points, Offer Details).
Codified Execution: We inject those answers into massive, pre-engineered system prompts. We effectively 'hard-code' the Senior Marketer's brain into the request.
The 'Codified Expertise' Insight
The real value isn't the AI generation; it's the guidance.
It’s like sitting down with a Senior Copywriter who says: 'Don't worry about the writing. Just tell me who we are talking to and what we want them to do. Oh, there's a discount? Let's do a back-to-school message and mention the queue-skipping bonus.'
We captured that specific 'briefing conversation' and turned it into code. The user provides the raw facts; the system provides the marketing degree.
Ready to fix the engine?
Let's find a time for a 30-minute chat about your plans and hopes.
Maybe I can help, or I can direct you to someone who can.
Based in Kranj, Slovenia. Operating Globally.