HOW I GOT HERE.
I'm Misha Manko - a Ukrainian technologist working independently across research, analytics, and software. I spent the last decade in the weeds of technical SEO, schema markup, and content infrastructure for businesses in regulated industries: law firms, medical practices, finance companies. Businesses where a single mis-rendered page or a broken schema field can cost six figures in lost traffic.
What I learned in that decade is that most of the "best practices" people repeat online fall apart the moment you pull real data. Google's documentation says one thing. The logs say another. The difference between those two things is where the interesting work lives.
ALWAYS HAVE.
WHAT I'M BUILDING NOW.
Three active projects, each independent, each with different economics. The thread connecting them is the same: I want to own the thing I'm working on. Not just bill hours on someone else's problem.
AI Visibility research. I operate a 47-site network with instrumented bot logs, publish findings on how AI models actually read websites, and audit client sites against that research. This is the flagship public offer - productized as a $2,500 fixed-price audit with a clear deliverable and timeline.
Shopify apps and iOS apps. Small, opinionated software built around gaps I've seen personally - either as a store operator, a phone user, or a developer working with these platforms. Different audience than the research work but the same philosophy: pragmatic, owned, transparent.
Consulting. For businesses where the audit isn't enough - multi-site networks, complex content infrastructure, enterprise AI visibility roadmaps. Custom-scoped engagements, 3-6 months, fixed scope and fixed price. This is where my research work applies directly to bigger strategic problems.
WHAT I DON'T DO.
I don't run a traditional agency. I don't do retainers without clear deliverables. I don't promise AI citation rankings - nobody can. I don't take on projects where I'd be a generalist among specialists. The list of things I decline is much longer than the list of things I accept, and that's on purpose.
I also don't chase the AI hype cycle. Most of what's labeled "AI-first" or "AI-powered" is either bolted on to make copy sound current or applied to problems that didn't need AI in the first place. My stance is simple: AI is a tool, sometimes the right one, often not. Measurement before enthusiasm.
THE SHORT VERSION.
I read the logs. I build the thing. I ship it. Then I write about what I found, so the next person doesn't have to repeat the work. If that sounds useful to you, we should probably talk.