I work in enterprise technology and build my own AI systems. I have 20+ years of experience in software development, testing and product delivery, mostly on large banking and payment platforms. Today I lead software development teams delivering cross-border payment platforms under strict regulation and constant audit.
On the AI side, I build and operate my own products end to end, using open-weight language models, data pipelines, retrieval, APIs and automated checks on my own infrastructure. AI speeds up the coding; I decide what is worth building, how the system should work and what counts as a correct result.
My products are running in production: Signal, a daily news briefing with video episodes on YouTube; a blockchain explorer with sanctions screening; a legal research assistant; and quantised open-weight models downloaded thousands of times. artjoms.ai is my sandbox for new technology.
My AI projects are independent of my employer, and the views here are my own.
Top skills
Product management · LLMs · RAG · model quantisation · AI solution delivery
More skills
Business process analysis · automation · AI governance · LLM benchmarks
The websites are only the visible interface. Behind each one is a system I built and run myself — in use, not on slides.
Blockchain explorer
btc.artjoms.ai
Check a crypto wallet before you take money from it. Bitcoin — 678 million addresses, grouped by co-spend analysis into 117 million likely wallets — plus USDT and USDC on TRON and Ethereum, against sanctions lists and issuer freezes. Updated around the clock.
Signal, not noise. Every day the pipeline reads hundreds of news items on tech and AI, markets and geopolitics, law and regulation, and keeps what matters: what happened, what changed and why. Daily briefings and weekly reviews in English and Russian, plus video episodes on YouTube.
Free · daily · English and Russian · youtube.com/@artjomsai
Ask about tax, startup or sanctions law and get an answer with the exact article it is based on — anything beyond the law texts is marked. Switzerland, the EU, the UK, the US, Asia and the UAE. A research tool, not legal advice.
Today I didn't follow the news. I asked Claude to check what happened on the blockchain, using only the API of my own blockchain explorer (see my previous article), and to write the story for me.
All of it is based on blockchain data that I store myself and share through an open API. Claude did a very nice job – one more example of what blockchain data and AI can do together.
What the data shows – transparency and control:
The good: Tether and Circle froze $2.2 billion of the stolen USDT and USDC.
The bad: $729 million moved on to addresses we can't follow. For some wallets the blacklisting came just 36 seconds too late.
The ugly: 18,270 BTC sit untouched in 19 wallets. Everyone can see them, nobody can freeze them.
Besides the infographics, Claude also found a few bugs in the public API of btc.artjoms.ai. The main one is already fixed, so the explorer is more robust now.
Important: my explorer covers only Bitcoin, plus USDT and USDC on Ethereum and TRON.
BTC.ARTJOMS.AI ON-CHAIN TRACE · BITCOIN · ETHEREUM · TRON Binance hack: what was taken and what was stopped From 20:20 UTC on 29 Oct 2026 Binance hot wallets on three chains sent their funds to fresh addresses.
Is it possible to vibe-code a blockchain explorer in 2026?
I got the idea for this article one summer evening, on vacation by the sea. I had started a vibe coding project and had already spent a couple of evenings on it. It's a blockchain explorer.
This morning my news pipeline flagged a pattern: the biggest AI companies are pouring billions into forward-deployed engineering. Old habit: I fact-checked it.