$ marketing-engineer instance, proposed for deployment to Chainstack
// status
// modules
User research with US-based and English-speaking audiences, using both qualitative and quantitative methods.
Uploadcare, a file-infra SaaS in the same audience class as ChainstackHands-on operator across paid channels in B2C and B2B, with a working view of when paid is the right primary lever and when it's a layer on top of something else.
MEL Science: $50K/mo solo on Google, Pinterest, Reddit. CAC −12%, CTR +18% via systematic A/B and creative cycles.Uploadcare: ran experiments across Google, Reddit, Quora, Carbon Ads, LinkedIn on smaller budgets. The takeaway: for a developer audience, paid works as a layer on top of content and education, not as a standalone acquisition channel.Daily LLM use across product, marketing, and design work. New tools earn a place in my workflow when they ship better outcomes than the previous setup.
Claude Code with MCP Figma on a current product-design project, building a design system and prototypes that convert directly into frontend code.n8n agentic content generation on a previous project. The pipeline shipped social posts; the project is closed, and today I'd architect it differently with current models.Cowork over a few iterations.I like making data make sense. For a developer audience, web analytics is always partly broken (adblockers, multi-context, multi-account), so making sense often means modeled rather than measured.
Uploadcare, contributing to the company's acquisition by Tiugo.Stronger on structure than on prose, which is why AI fits naturally into my writing workflow. I hand it the structure and the facts, and it does the writing.
Comfortable on low-code platforms, and able to dig into a custom PHP file in WordPress when needed. This page is a recent example, built in Cowork and hosted on Vercel.
// changelog
// about
I like SMB consulting and running a bunch of projects in parallel. What I keep running into is FOMO: a non-tech owner rarely actually wants to optimize their infrastructure, while AI implementation success stories are everywhere. So when Losha pinged me about Chainstack, the appeal was obvious. This is a real polygon for AI workflow experiments at scale.
On Web3 specifically, I'm curious and convinced by SheFi. Foundational on RPC, wallets, staking, on-chain primitives; not protocol-engineer-deep. What I'd bring instead is fast technical-domain ramp-up (research chemists ramp on infrastructure topics quickly) and the fact that dev-tools SaaS marketing is the job I've already done. The protocols are different; the marketing problem isn't.
Losha and I worked together at Inita. We ran a few tests and tuned local conversions. When he referred me, the shorthand was probably "orchestra-type, no-bullshit".
Telegram is the fastest way to reach me. Email and LinkedIn also work.