Modernise legacy systems
Established applications are renewed step by step. User accounts, data, permissions, interfaces, and important business processes are preserved wherever possible.
westudiolab
Modernise legacy systems. Preserve connections. Build lean applications.
We evolve established software into modern systems without carelessly abandoning user accounts, data, interfaces, or proven workflows. The result is lean websites, web apps, and AI-assisted applications that fit the business and its existing connections.
01 / Services[01] [02] [03]
From a tangled legacy system to a new web app
Not every project needs a large systems audit. We distinguish between complex modernisation, a new digital application, and targeted AI integration – and choose an entry point that fits the actual task.
Develop websites & web apps
From a focused company website to a custom web app: lean in concept, centrally maintained, multilingual, and designed around genuine user value.
Integrate AI with purpose
We develop AI-assisted tools and automation where they genuinely improve processes – with understandable data flows and without artificial AI washing.
Strategic Initial Consultation (Audit)
Strategic Legacy Analysis – by agreement
For complex or business-critical systems, we first clarify architecture, dependencies, risks, and realistic modernisation paths. Scope and fee depend on the actual complexity – from a clearly defined system to an established enterprise landscape.
\"The result is an understandable baseline assessment with prioritised next steps – not automatically an oversized report.\"
Request Technical Assessment
By agreement, the analysis fee can be credited in full or in part when a subsequent project is commissioned.
From the Lab
LAB EXPERIMENT 001
Swarm Moiré – Collaboration Becomes Visible
This experiment visualizes how 2500 autonomous agents form a collective movement from local decisions. The swarm is not a decorative effect, but a visible excerpt of our work on multi-agent systems.
Open experiment →Selected Reference
Richard Dragulejev Photography
A visually focused multilingual portfolio for photographer Richard Dragulejev – with clear image direction, centrally maintained content, and automated English translation.
View website →02 / Expertise
Decades of systems experience meet applied artificial intelligence
westudiolab is the technology and development division of WE Studio Media GmbH. It grew out of many years of work with software, digital production processes, and artificial intelligence. While WE Studio provides spaces and photo and video productions, the Lab develops digital systems, web apps, and AI-assisted tools – independently, but under one company roof.
Peter Gesierich
Founder & Lead Architect
40+ years of software development – from punch card systems and the Sinclair ZX Spectrum to today's web and AI infrastructure. Specialist in systems integration, pragmatic AI orchestration, and technical search optimisation. The focus is on understandable solutions that evolve established systems step by step while protecting business-critical connections. Authorised signatory of WE Studio Media GmbH and responsible for building its new technology division, westudiolab.
David Reger
DATA & AI PLATFORM ENGINEER
Specialist for data platforms, AI integrations, and scalable system architectures. Focus on building the data foundation for productive AI applications – from data integration and RAG systems to agentic workflows. Expertise in Python (FastAPI, SQLAlchemy), SQL/database engines, Kubernetes, and modern data stacks. Experience with web scraping, PDF parsing, ERP integrations, semantic search, and event-driven architectures. Technical focus: Lakehouse platforms, streaming systems, data processing, LLM orchestration, and production-ready AI deployments. From data sourcing to the robust deployment of AI-enabled services.
03 / Product & System Strategy
Evolve what exists. Build new products on solid foundations.
Strategy does not begin with a programming language. We consider goals, users, workflows, and existing value, then shape a solution that works today and can continue to grow tomorrow.
[01]
Purpose and value first
We clarify which problem the application should solve, who will use it, and which workflows genuinely need improvement. Only then do we decide on features and technology.
[02]
Preserve what matters
In established systems, we protect user accounts, data, interfaces, and working processes. For us, modernisation means thoughtful evolution rather than a premature complete rebuild.
[03]
Build new products lean
New websites, web apps, and AI applications are designed to be modular and extensible. The architecture follows the project rather than a preferred programming language.
[04]
Grow step by step
We deliver verifiable stages, test against real requirements, and expand only where measurable value emerges. This keeps the product understandable, manageable, and sustainable.
Technical process for legacy modernisation
For technical decision-makers, this is our proven five-phase model. It keeps change controlled, reversible, and as unobtrusive as possible for existing users.
[01]
Phase 1 – Visibility & Security
Mapping the system landscape, processes, and dependencies. Establishing monitoring and blue-green deployment strategies to ensure every change remains reversible.
[02]
Phase 2 – Modularise, Don't Destroy
Structuring existing code into a "Modular Monolith". Introducing clean APIs between modules instead of "Wild West" data access.
[03]
Phase 3 – Strangler Approach & Facade Architecture
Implementing the Strangler Fig Pattern: New features are built externally, legacy endpoints are replaced incrementally without users noticing the transition.
[04]
Phase 4 – Data Migration Without Shock
Utilising techniques like Change Data Capture and read replicas. Parallel operation of old and new data paths with constant validation.
[05]
Phase 5 – Decommission & Cleanup
Targeted shutdown of legacy components and removal of dead code. Handover of a lean, modern core system to your team.
04 / Technology & Data Sovereignty
Technology follows the project – not the other way around.
We combine proven software engineering with the purposeful use of artificial intelligence. The goal is not the longest technology list, but a solution that remains understandable, maintainable, and under your control.
[01]
Websites 3.0
A digital presence becomes an intelligent application: it connects content, data, and processes, assists users with AI, and can automate useful workflows without surrendering control or data sovereignty.
[02]
Architecture that fits
We selectively evolve existing technologies or choose a stack for new products that fits usage, team, and growth. What matters is not the programming language, but a durable and maintainable solution.
[03]
Controlled infrastructure
Our own AI infrastructure on NVIDIA DGX Spark and a Hetzner server location in Nuremberg enable controlled, privacy-conscious operating models. Depending on the project, sensitive data can be processed locally or in Germany.
Technology in detail
For technical teams: the concrete selection always follows the task, the existing environment, and the preferred operating model.
[01]
Backends & APIs: Python (FastAPI, SQLAlchemy, Pydantic), Node.js (Next.js, Express), REST APIs, GraphQL, and WebSockets for real-time integration.
[02]
AI & LLMs: RAG pipelines, agentic workflows, multi-model orchestration (OpenAI, Anthropic, and local models via Ollama), LiteLLM Gateway, semantic search, and embeddings.
[03]
Data & integration: relational and document databases, lakehouse architectures, web data extraction with Playwright and BeautifulSoup, PDF parsing, ERP/CRM integration, and event streaming with Kafka.
[04]
Frontend & SEO: Next.js with App Router and Server Components, React, Tailwind CSS, technical SEO, structured data, accessibility, and Core Web Vitals.
[05]
DevOps & infrastructure: Docker, Kubernetes, CI/CD with GitHub Actions, Ubuntu Linux, Caddy reverse proxy, systemd, monitoring, NVIDIA DGX Spark with Ubuntu-based DGX OS for local AI workloads, and Hetzner hosting in Nuremberg.
Insights from the Lab
Using local AI where data needs to remain protected
Not every AI application belongs in a public cloud without careful consideration. At westudiolab, we evaluate local models and clearly defined interfaces for tasks where control over data, cost, and operations matters.
The aim is not to use the largest possible technology, but to build a solution that fits the real process and whose limitations remain understandable.
Modernising legacy systems without losing existing users
A clean-sheet rebuild may sound simple, but it often puts accounts, permissions, data relationships, and familiar workflows at risk. We therefore begin by understanding real dependencies and replacing parts step by step.
This allows a new, lean system to emerge while important connections are transferred and verified in a controlled way.
Central content instead of separately maintained language versions
Multilingual websites should not consist of separate text copies that quickly drift apart. Our content models begin with one centrally maintained German version; the English version is generated automatically and can be reviewed editorially.
We apply the same principle to new web apps: one reliable source, multiple suitable outputs.
Ready for the Future?
Briefly describe your existing system or idea. We will respond with an honest assessment of the most useful next step.