S/01
Architecture Assessment
Find bottlenecks, risks, coupling, failure modes, and delivery blockers before they reach production.
EU-based · Open to advisory & fractional work
SOFTWARE & AI SYSTEMS ARCHITECT
Resilient platforms,
shipped under real load.
I architect multi-region, high-load systems — and accelerate delivery with AI agents, with a human owning every decision.
30-min call · no pitch · a concrete next step you can act on.
- 14+ yrs in software
- Multi-region platforms
- High-throughput systems
- Led engineering teams
TRUSTED ACROSS
PlaysonPlaytikaDMND.games
// WHAT I’M KNOWN FOR
Systems where latency, consistency, and failover are non-negotiable — and governance is designed in, not bolted on.
01Multi-region failover
active-active · geo-routing · Cloudflare
02Event-driven platforms
Kafka · CQRS · event sourcing
03High-throughput systems
low latency · high availability
04Cloud infrastructure
AWS · Kubernetes · Terraform · FluxCD
05Observability
Datadog · Grafana · SLOs
06AI-assisted delivery
agentic workflows · OpenSpec
07Engineering leadership
teams up to 20+
08Security by design
IRSA · ABAC · secure CI/CD
// 01 — SERVICES
Architecture and execution support for complex software platforms.
S/02
Multi-region Platform Design
Design failover, traffic steering, consistency strategies, observability, and safe operations across regions.
S/03
AI-Assisted Delivery
Build agentic workflows for planning, coding, reviews, testing, documentation, and automation.
S/04
Cloud & DevOps Review
Improve Kubernetes, AWS, Terraform, CI/CD, security posture, monitoring, and cost efficiency.
S/05
Fractional Architect
Hands-on architecture leadership for teams that need senior technical direction.
S/06
Engineering Coaching
Help engineers unblock complex technical decisions and raise system design quality.
// 02 — SELECTED WORK
Distributed systems that had to stay up.
Playson
iGaming · PlatformMulti-region iGaming platform modernization
Contributed to the evolution from a C++ monolith into a large microservices platform — geo-routing, failover flows, cross-region context transfer, consistency strategies, observability, and production ownership.
DMND.games
Startup · Real-timeReal-time game backend and cloud infrastructure
Designed and operated backend services, cloud infrastructure, monitoring, alerting, and abuse / DDoS response for a fast-moving startup environment.
Playtika
Social casino · High-loadHigh-load social casino product features
Delivered backend features for Caesars Casino — event-driven high-load product mechanics and production stability work on the Java / Spring Boot ecosystem.
Open Source
Kafka · node-rdkafkaKafka cooperative rebalance contribution
Contributed Cooperative Rebalance support (PR #1081) to node-rdkafka, reinforcing deep Kafka and distributed-systems expertise.
// 03 — METHOD
From unclear technical risk to production-ready architecture.
- 01Diagnose
Read the codebase, traffic, and incidents to find the real blockers.
- 02Model
Model load, data flows, consistency, and the failure modes that bite.
- 03Design
Resilient architecture with trade-offs written down, not implied.
- 04Ship
Incremental rollout behind flags, with a tested rollback path.
- 05Observe
SLO dashboards and Datadog / Grafana monitors that trigger real action.
- 06Automate with AI
Encode delivery into agentic, repeatable workflows.
// 04 — AI ENGINEERING
AI-assisted delivery, owned by a human architect.
I use AI agents to take on the unglamorous 80% of delivery — planning, implementation, review, tests, and docs — so senior time goes to the decisions that actually need judgment. The pipeline below is how that runs.
AGENTIC DELIVERY PIPELINE
Plan mode
Claude · Cursor
Role workflows
Subagents
OpenSpec
Change mgmt
Agent APIs
Automation
// 05 — TECH DEPTH
The distributed-systems stack I run in production.
The layers I design, own, and operate — follow the signal from the edge down to security.
REFERENCE ARCHITECTURE · LAYERS I OWNrequest flow
L1Edge & Routing
Get every user to the nearest healthy region — and steer traffic away from failure before they notice.
L2Compute & Orchestration
Reproducible, self-healing infrastructure defined as code — multi-AZ by default.
L3Services & APIs
Loosely-coupled services behind typed, versioned contracts.
L4Eventing & Data
An event-driven core with replayable state and deliberate consistency between services.
L5Observability
SLO-driven signals and dashboards tuned to catch real incidents early.
L6Security & Governance
Least-privilege access and a verifiable software supply chain, with governance built in from the first diagram.
CROSS-CUTTING · LANGUAGES
CROSS-CUTTING · AI ENGINEERING
// 06 — ABOUT
Hands-on architect with research roots and production scars.
I started in science, building sensors and the data-acquisition systems that read them on bare-metal microcontrollers. Using DSP, I pulled correlations out of raw electrical and magnetic signals and turned them into metrics that real-world automation systems could act and decide on. That work set a standard I still hold: take something noisy and physical and make it trustworthy enough to bet a decision on.
Since then I've built across most of the stack and most of the industry, from early-stage startups to established product companies — mobile and desktop apps, frontend, high-load backend platforms, games, and the big-data pipelines and CI/CD behind them. Over time my focus narrowed to the harder end: geo-distributed clusters with live replication and cross-region user-data transfer, high-reliability event-driven systems on Kafka, and complex cloud-native products taken from zero to production with AI agents doing real work under my direction.
What keeps me in it is the difficulty. I'd rather learn a better approach than defend a familiar one, so I stay current on tooling and use it to ship quality at speed. My priority is getting the design right the first time, so the business can grow on it without a rewrite when scale or requirements change.
- 14+
- years in software
- 20+
- engineers led
- 7.5k+
- tx/sec
- EU
- based & available
Have a platform, delivery process, or AI workflow that needs senior eyes?
In 30 minutes we'll pressure-test your architecture or delivery bottleneck and leave you with 2–3 specific, prioritized moves — whether or not we work together.
Typically responds within 24h · EU business hours