I'm Dzmitry Harupa, Head of Architecture. 17 years of high-load systems, 10 of them in iGaming. I design architecture, take AI/ML to production and still write code myself. TOGAF, SAFe and AWS certified. 100% remote.

From a one-off consultation to a fractional CTO role — pick the depth you need now.
Architecture reviews and roadmap sessions. One call — specific recommendations.
From concept to a working prototype. Architecture documentation is part of the result.
Technology leadership part-time: from monthly advisory to a full CTO roadmap.
Hands-on workshops for your team: architecture, AI/ML, working with ADRs.
Enterprise architecture, AI/ML and ten years of iGaming domain knowledge — hands-on, in code.
Senior-level expertise without a full-time salary and overhead.
The scope changes with your needs. You pay for hours, not a seat.
Hiring a CTO takes months. Here the first call is this week.
Start with a free consultation. A retainer comes after you see results.
Working prototypes, executable architectures, code in production.
Flexible timezone coverage and async communication.
>
200K+ msgs/sec across 1000+ systems, <100ms p99 latency.
40% of transactional load migrated; release lead time down 57%, deploys under 15 minutes.
Decisions on data. Results in numbers.

Ten years in iGaming: Casino, Sports, Poker, PAM, payments and CRM. Production AI/ML on AWS Bedrock, SageMaker and LangGraph.
I grew an architecture practice from 6 to 12 architects plus 20 system analysts, mentored three of them to Head-of-Technology roles and led 200+ engineers. I still write production code.
What I work with: architecture, AI/ML, iGaming.
The recursive-context story continues: four releases in one day teach the agent a persistent code graph (Graphify) under strict distrust rules — freshness gate, facts only from live files. Plus a controlled A/B test of graph vs grep on NestJS: −71% tokens on subsystem understanding, parity elsewhere — and the hidden sub-agent costs that make naive benchmarks lie.
ReadFrom MIT's Recursive Language Models paper to the recursive-context skill in one day: why a huge context window doesn't solve big data, how a fan-out of isolated sub-agents with honest coverage replaced a Python library, and what only live runs could catch — from a baseline that turned out too good to a 14-millisecond bug.
ReadA picture-diagram checks nothing. Event Storming does: every event has a cause, every command an effect. How an agent runs a board through a deterministic sequence analyzer, closes mechanical gaps to zero, and leaves unresolved business questions visible. On a live Forklift project — 5 contexts, 17 gaps caught.
ReadDescribe the problem in a couple of paragraphs — or skip ahead and book a free 30-minute call.
calendly.com/d7561985 · no commitment