Volker WidorMfA²ISoftware architecture, custom development and applied AI
Bring data together and process it efficiently.
Automate recurring tasks.
Bring new ideas to life.
I combine extensive experience in software architecture and development with practical AI use. Together, we clarify your requirements and design the solution. I then take care of the technical implementation.
I consider the affected processes and systems as a whole: from interfaces and additional features to the gradual, complete replacement of existing software products. I integrate AI where it suits the task.
Avoidable routines and manual handovers can tie up resources needed for your core business.
Software and AI.
Together, we clarify what you need. I then implement the right solution.
For companies with around 10–200 employees in Northern Germany – with or without an in-house software development team.
I work with business departments, users and IT to clarify what your software needs to do. I develop suitable applications and interfaces, connect existing systems and modernise software that has evolved over time. I automate suitable steps using conventional software or AI.
Existing applications can be extended or replaced in agreed stages, up to complete replacement. Running systems in parallel, comparing results and retaining a fallback option support a controlled transition.
Your outcomeDepending on the agreed scope: an evaluated pilot or a usable solution with documentation and an introduction for users. The aim is to reduce manual data transfers and unnecessary rework, creating more room for your core business.
I help your team use coding agents on a clearly defined, real development task. I combine my experience in software architecture, development and reviews with my own practical use of agents.
Together, we prepare project knowledge and define rules for tasks, architecture, tests and reviews. We explore which steps agents can take on and where additional checks are needed. We assess quality, rework and effort against agreed criteria.
Your outcomeA documented way of working, tested on the chosen development task, together with an assessment of its results and limitations. Your team can build on this and develop its use of agents independently.
A separately commissioned first step
If the right approach is still unclear, I examine a selected area of your operations or development work. The result is a set of prioritised use cases and a basis for deciding on a clearly defined first implementation step.
Clients, business departments, users and IT contribute their requirements and practical knowledge. I examine processes, data and systems and develop a sound concept. We agree the starting point and evaluation criteria before implementation.
We agree manageable steps with clear deliverables and verifiable results. I offer fixed prices for tasks with a sufficiently defined scope. For open questions or ongoing work, we bill for time spent within an agreed budget.
I take care of the agreed technical implementation. The users involved provide feedback and try out the results.
We review the results against the agreed criteria. Documentation and an introduction for users enable independent use. We tailor ongoing support to your organisation: I provide maintenance, support and further development within the agreed scope, working with your IT team or other partners where needed. Source code, configurations, tests and necessary access are handed over as agreed.
After introduction, I remain your point of contact for questions and issues in day-to-day use. When processes or requirements change, I adapt and develop the solution further. This allows it to support your business over the long term and create more room for your core work. We agree the scope and form of support to suit your needs.
In my development work, I use AI for suitable tasks – from architecture, tests and quality checks to preparing work for approval by the people responsible. Requirements, tests and reviews set the standard for results. Final assessment and responsibility remain with people.
Data protection and the handling of confidential information are part of the design from the outset. Together, we clarify which data is processed, how and where – for example, using local AI models or suitable external services. Final assessment and responsibility, including data protection and approvals, remain with people.
The industrial project illustrates my work on architecture, data integration and implementation within a team. Two applications of my own add current practical AI experience in development and operation.
In operational use during the trial phase
Starting point: An industrial company wanted to develop an existing AR prototype into a solution for everyday work. Previously, employees had to relate information from 2D drawings to physical parts and document inspection results manually.
My contribution: I took responsibility for the software architecture and substantial parts of the implementation, as well as the roles of Lead Developer and Scrum Master. A key focus was integrating different live systems for CAD management, material data and quality assurance. The solution was developed with the development team, in close consultation with manufacturing, quality assurance and IT.
Solution and benefits: An AR headset overlaid the components to be inspected onto the real working environment. Employees could record inspection results directly in digital form. Employee feedback and the available results indicated significantly faster processing, fewer matching errors and less need for follow-up checks.
Deployment status: When I left the project, the system was already in operational use and was still in its trial phase.
The complete workflow is used regularly in everyday work
Gathering information, preparing papers, tracking responses and documenting results: these tasks accompany many decision-making processes. An organisational solution of my own shows how software and AI can work together.
Starting point: In an organisation’s committee work, resolutions by circulation needed to be prepared and documented in a traceable way. This required bringing together rules, earlier resolutions, responses and deadlines.
My contribution: I designed and developed an end-to-end workflow – from describing the matter to preparing an email draft for review, reporting status and documenting the evaluation.
Solution and benefits: AI prepares the resolution text using the applicable rules and earlier examples. The user reviews and sends the complete email draft. Software and AI then track responses and deadlines, report progress and create a final PDF evaluation. This keeps preparation, consultation and the outcome connected and traceable.
The system uses a local language model and an approved LLM provider; protected email data is processed locally.
Organisational solution
From a request in a messaging app to a PDF evaluation.
User + AI
The user describes the proposed resolution in a messaging app.
AI drafts it using statutes, rules and earlier examples.
Placed in the drafts folder with all recipients and Reply-To.
User
In the email application, by the user
Software / AI
Text and deadline information sent by messaging app.
Ongoing evaluation and status updates by messaging app.
Outcome sent by messaging app, including a PDF evaluation.
Deployment status: Developed for my own organisational work; the complete workflow is used regularly in everyday practice.
In regular operational use in my own daily life for some time
Starting point: In my own energy system, charging needed to fit everyday routines, changing electricity prices and on-site solar generation. Charging targets, absences and household energy needs all had to be taken into account.
My contribution: I designed the processes and developed the planning and control software with AI assistance. I was responsible for integrating existing data sources, testing and validation in real operation.
Solution and benefits: The software aligns charging needs, target times and absences with dynamic electricity prices, solar forecasts, household consumption and battery storage. It selects suitable charging windows for solar energy and low-cost grid electricity and adjusts the plan as conditions change.
Deployment status: A personal project that has been in regular operational use in daily life for some time.
Software architecture and custom development
Over many years as a freelance professional, I have designed, developed and modernised software for a range of specialist fields. My responsibilities have included clarifying business requirements, software architecture, data models and interfaces, as well as implementation and testing. As Lead Developer and Scrum Master, I have also coordinated development work within teams.
With a German Diplom in mathematics and computer science as a subsidiary subject, I work through complex tasks and develop sound solutions. For me, technical relationships and the user’s perspective belong together: software needs to fit the actual tasks and be understandable in everyday use.
I complement this experience with practical work using coding agents and AI applications. Solutions of my own that I use regularly show how I apply AI in development and operation and check its results.
MfA²I
Manufactory for Applied Artificial Intelligence
The name stands for applied AI and the care of a manufactory: thoughtfully designed software that fits your company’s tasks and processes.
For my clients, this means combining an understanding of their work with technical implementation. In discussion with clients, business departments, users and IT, I clarify needs and develop a suitable concept. My experience with established applications, data models and interfaces helps me integrate the solution into existing processes and systems. I support the journey from the initial question to a tested result and continue to help during day-to-day use. What matters is that the solution eases the workload in practice and frees up capacity for the company’s core tasks.
My focus is on companies with around 10–200 employees in Northern Germany. I am based in Kiel; initially, on-site collaboration is planned within roughly four hours’ driving distance.
Briefly describe your starting point – a workflow, a development task or your wish to use AI for a specific purpose. You do not need a fully defined use case.
volker@widor.biz