Services

How I can help

  1. Understand.The goal, the system’s surroundings and the rules the old system quietly follows.
  2. Design.The architecture, the decisions and an approach the team and the AI agents can sustain.
  3. Verify.Tests, evals and measurements, not “the AI checked it”.

I cover the whole workflow – from design and analysis through development to deployment and operations. Faster and cheaper than before thanks to AI, with quality control that comes from tests and measurements, not impressions. I price individually after the intro call, always for the result.

01

Architecture consulting and reviews

Companies facing an important technical decision or needing an independent view.

  • review of an architecture or a vendor’s proposal, including risks and a cost estimate
  • audit of code, database and technical debt
  • a second opinion: rewrite, modernise, or buy an off-the-shelf product?
What you get
A concise report with findings ranked by severity, concrete recommendations and a presentation for management and IT.
Typical first step
Send me the materials (design, vendor proposal, code access) – you usually get the assessment within 1–2 weeks, depending on scope.
02

Custom systems

Companies for whom spreadsheets and boxed software are no longer enough – from smaller firms to in-house IT of large organisations.

  • internal information systems, records, workflows and approvals
  • web applications with sign-in via company accounts (Microsoft, Google, SSO)
  • deployment to the cloud (Azure) or on-premises, CI/CD, monitoring and operations
What you get
Analysis → design → development in stages with a demo of each stage → deployment and operations. The code and documentation belong to you.
Experience
Five years of in-house development at VZP ČR (application modernisation, team lead), systems for banks and financial companies, Job Scout, Payment matching with AI and other custom applications – see Projects.
Typical first step
An initial analysis and design of the first stage.
03

Modernising legacy systems

Companies with a key system that is 10–25 years old whose modernisation they kept postponing because it looked expensive and risky.

  • quick diagnosis: current state, risks, recommended strategy
  • analysis and design: inventory, business rules extracted from code with the help of AI, target architecture, roadmap and estimate
  • gradual migration with tests that check the new system behaves exactly like the old one
What you get
A clear plan for the system – what to keep, what to modernise, in which order and for how much.
Experience
I spent years building systems in Delphi and PL/SQL (MONETA, Raiffeisenbank, ARBES); today I design solutions in .NET and Azure. I understand both worlds.
Typical first step
A diagnosis in one week. If you then order the analysis, the diagnosis fee is deducted.
04

Integrations and APIs

Companies where data is re-typed between systems or integrations keep failing.

  • integration with partners, carriers, banks, government and internal systems
  • API design, message queues (Azure Service Bus), security and secrets management
  • monitoring, so you learn about a failure before a customer reports it
What you get
A reliable connection with documentation and visibility into what is happening.
Experience
Integrations with carriers (PPL, DPD, Czech Post) and partners on a smart parcel locker platform, internal system integration at VZP.
Typical first step
An integration design and estimate.
05

AI in your solutions

Companies that want to use AI fully and securely – neither avoiding it nor deploying it blindly.

  • where AI makes sense in your system and where it does not
  • assistants and agents over company data (LLMs via API, tool calling, document search)
  • quality measurement, protection against prompt injection and abuse, cost control
What you get
Use cases, architecture, risks and an estimate of running costs – or a working solution straight away.
Experience
Job Scout – an AI agent in production, with every call and its cost recorded and limits against abuse. Payment matching with AI – the model decides only the unclear payments, code checks every answer it gives and evals in CI guard the accuracy.
Typical first step
An AI adoption proposal in 1–2 weeks.
06

Architect and team reinforcement

Teams lacking an experienced person for architecture, technical decisions or the move to AI-assisted development.

  • architecture and technical decisions recorded so you can come back to them (ADRs)
  • code reviews, mentoring, CI/CD and rules for working safely with AI agents
  • communication with management and the business in plain language
What you get
Long-term on contract, part-time or full-time. For the right team, employment too.
Experience
Leading a development team at VZP ČR reporting directly to the IT director, architecture of a smart parcel locker microservice platform.
Typical first step
An intro meeting with the team.

Not sure which of these you need?

That’s fine – we will find out on the intro call.

Book a free call