Software architecture
Systems shaped around how the business actually works, with clear domain models and boundaries that still make sense in three years.
Domain-driven design · Event-driven systems · Technical roadmaps
Precision. Delivered.
Software built to last, for platforms that can't stand still.
Tailored software, architecture and applied AI for business-critical platforms. Senior engineering, from first sketch to production.
Design the system, then build it.
From the first architecture sketch to the code that runs in production.
Applied AI
LLMs and agents built into existing products and processes, connected to your systems through MCP. Reliable, measurable and affordable to run.
LLM integration · MCP servers · Agents
Platform renewal
Legacy applications brought up to date step by step, while the business keeps running.
Symfony LTS · PHP 8.x · Refactoring
Performance & scale
Message queues, workers and finding the real bottleneck under production load.
Symfony Messenger · Workers · Profiling
Human insightYour people, processes and know-howArchitectureStructure that keeps it reliableAI systemsAgents and automation in production
AI integrations, done properly.
A core discipline at Orbis: AI that works inside your existing platform, stays within budget and holds up in production.
- 01Standalone agentsAssistants for support, intake and back-office work, connected to your systems through MCP, with clear boundaries and a human in the loop.
- 02Coding agentsAI-assisted development set up for your codebase, conventions and CI.
- 03Review agentsAutomated review of code, content or documents before a person signs off.
- 04MonitoringLogging, evaluations and alerts, so you see what the AI does and how well.
- 05Token & cost managementCaching, model routing and budgets that keep running costs predictable.
- 06AI strategyWhere AI pays off, what to build first, and how to stay compliant with GDPR and the EU AI Act.
06AI strategyUse cases, priorities, compliance
01Standalone agentsSupport, intake, back office
02Coding agentsCodebase, conventions, CI
03Review agentsCode, content, documents
MCP connectorsA standard, secure bridge between agents and your systems
Your platform & dataAPIs, databases, workflows
04Monitoring
05Token & cost
Agents connect to your existing platform through MCP. Monitoring and cost control run across every layer.
Platforms that keep running.
Releases
Zero-downtime delivery
Upgrades, migrations and new features go live without interrupting your users or operations.
Engineering
Scale
Built for peak load
Queues, workers and caching that hold up when traffic and workload spike.
Performance
AI
AI that earns its place
Integrations measured on real data: time saved, errors avoided, cost per run.
AI integrations
A senior practice, deliberately small.
Orbis works directly with the people who own the platform. No account managers, no unnecessary overhead, no handovers: the engineer who designs your system is the one who builds it.
Orbis is led by Niels de Klerk, based in the Netherlands and working with clients internationally. He holds a Master of Business Administration with advanced studies in management, and spent more than ten years in IT consultancy, project management and information analysis before moving into hands-on engineering over eight years ago.
That combination shapes every engagement. Technical decisions are made with the organisation, its risks and its goals in view, drawing on experience across e-commerce, logistics, healthcare, government, and insurance and financial services.
Change without the risk.
A new architecture, a framework upgrade or an AI feature: every change reaches production the same way.
- Small stepsEvery change is reviewable and reversible.
- VerifiedAutomated tests run green before anything moves on.
- MeasuredPerformance and AI output checked against real data.
Technical foundation.
The tools and practices behind every engagement, chosen for stability first.
Backend
- PHP 8.x
- Symfony 6.4 & 7 LTS
- Doctrine ORM
- REST & integration APIs
Async & scale
- Symfony Messenger
- Message queues & workers
- Caching strategies
- Profiling under load
AI integrations
- LLM APIs
- Retrieval & embeddings
- MCP servers & tool use
- Evaluation on real data
DevOps & quality
- AWS
- CI/CD pipelines
- Containers & infrastructure as code
- Logging & monitoring
- Automated testing
- Code review & documentation
Built to outlast the engagement. Code lives in your repositories, follows framework conventions and is fully documented, so it keeps running and stays maintainable long after delivery.
From first call to production.
1
Introduction
We talk through the platform, the goals and what success looks like.
2
Assessment
Architecture, code and pipeline reviewed, including where AI could help.
3
Plan
A phased proposal with priorities, risks and a realistic timeline.
4
Delivery
Small, tested steps all the way to production.
Precision. Delivered.
Let's talk about what's next.
Whether it is an idea or a running platform: thirty minutes, no obligation, and you leave with a clear next step.