Custom software for enterprises

Custom systems and integrations for business.

Process analysis, architecture, implementation and deployment. The system is built around how the business actually works; code, documentation and infrastructure stay on the client side.

Discuss a project See the services
  • Work in stages, from analysis to deployment
  • Integrations and process automation
  • Code and documentation handed to the client

Services

Four stages, four concrete deliverables

Every stage can be commissioned separately. At the end, you have a concrete deliverable and a decision whether to continue.

Analytical consulting

Before a single line of code exists, someone has to establish what the system must actually do — and what it must deliberately not do. I break the business process down, surface contradictory requirements, and point out where a custom build simply will not pay for itself.

  • Workshops and stakeholder interviews
  • Process and data-flow mapping
  • Structured functional and non-functional requirements
  • Feasibility study and risk analysis
  • Staged cost estimate
  • discovery
  • requirements analysis
  • feasibility
  • estimation

Architecture design

I design architecture for the actual load and the team that will maintain the system. The reasoning behind each significant decision is recorded.

  • Domain model and decomposition into modules or services
  • Integration design against existing systems
  • Technology and architecture selection, with the cost case spelled out
  • Performance, scalability, security, GDPR
  • Architecture decision records and C4 diagrams
  • system architecture
  • DDD
  • integrations
  • ADR / C4

Implementation

Code is delivered in short iterations, each ending with something you can open and comment on. There is no "we will show you everything at the end" phase — that is exactly where the most expensive misunderstandings come from.

  • Working increments every 1–2 weeks on a test environment
  • Unit, integration and end-to-end tests
  • Code review and static quality analysis
  • CI/CD pipeline from week one
  • Repository and backlog on the client side from day one
  • B2B web applications and Windows desktop clients for shop-floor workstations
  • software development
  • web application development
  • automated testing
  • CI/CD
  • code review

Deployment and maintenance

Deployment is not "uploading it to a server". It is migrating the data, moving people onto a new way of working, and making sure that when something starts to fail, someone finds out before the users do.

  • Data migration from legacy systems
  • Production rollout with a rollback plan
  • Monitoring, structured logging and alerting
  • Training for end users and administrators
  • Post-deployment documentation and handover
  • Optional maintenance agreement with response times
  • go-live
  • data migration
  • monitoring
  • maintenance

When this fits

When an off-the-shelf tool is no longer enough

Usually when the problem sits between processes, data and existing systems.

Data moves by hand

Information is re-entered between spreadsheets, warehouse systems or partner systems.

The process is specific

A packaged product needs expensive workarounds or changes a way of working that matters to the business.

Control needs to return in-house

The company needs predictable maintenance, access to code or deployment on its own infrastructure.

Licence independence

Lower licence costs, more control

Where it makes business sense, I choose components without licence fees charged per user, core or transaction. Before a migration, we compare the cost of change with the cost of keeping the current solution.

Microsoft Windows Server Linux — Debian, Ubuntu Server, RHEL-compatibleno CAL licences or per-core fees; support matched to the need
Microsoft SQL Server PostgreSQL (+ TimescaleDB for time series)schema, data and stored-procedure migration included in the rollout
Oracle Database PostgreSQLa possible alternative after assessing fit, migration cost and risk
Power BI Grafana · Metabase · Apache Supersetdashboards and reports without a fee per additional viewer
AWS and other hyperscalers Your own infrastructure — on-premise or private cloudmore control over cost, with responsibility for maintenance retained in-house
… and the next subscription Openly licensed componentsthe list is open-ended — during analysis we go through every recurring fee you currently pay

Deployment under client control

The system can run in your server room or private cloud. Any processing of data outside the organisation is defined consciously during design.

Integrations without per-transaction fees

Data exchange can run on purpose-built API services instead of a platform priced by transaction volume. Cost then follows integration complexity and infrastructure needs.

Code and documentation on the client side

The repository, documentation and infrastructure configuration can live in the client organisation from the start. The scope of rights transfer is set in the contract.

How the work runs

From the first conversation to production

The sequence is the same regardless of project size. Only the length of each stage changes.

  1. First meeting

    A free discovery conversation in which we establish together whether a custom solution makes sense in this case.

    deliverable: initial scope and a budget order of magnitude

  2. Analysis and requirements

    Workshops with the people who actually run the process. The output is a document describing what the system must do, where its boundaries are and what it integrates with — concrete enough to price the work and to compare competing offers.

    deliverable: requirements specification + staged estimate

  3. Architecture design

    System structure, data model, integration approach, technology choices and non-functional requirements. Every significant decision is recorded together with the alternatives that were rejected and why.

    deliverable: architecture documentation, C4 diagrams, ADRs

  4. Iterative implementation

    Short cycles, each ending with a demo of a working build. The client sets backlog priorities, so scope can be adjusted mid-flight without renegotiating the whole contract.

    deliverable: working increments every 1–2 weeks

  5. Production rollout

    Data migration, go-live, monitoring, training and a stabilisation period under close observation. Only after that do I consider the project closed.

    deliverable: system in production + full documentation

  6. Support and evolution

    Optional. Fixes, new features, dependency and security updates. Or the alternative: a complete handover to your in-house team with a supported transition period.

    deliverable: SLA agreement or handover to your team

Technology stack

Technology chosen to fit the problem, not the other way round

Technology follows the problem, the existing environment and the cost of maintenance — not fashion.

Applications and integrations

  • .NET / C#
  • Go
  • ASP.NET Core
  • REST / gRPC / GraphQL
  • RabbitMQ / MQTT

Data and reporting

  • PostgreSQL
  • TimescaleDB / ClickHouse
  • MS SQL / Oracle migrations
  • Grafana / Metabase

Interfaces

  • TypeScript
  • Vue / Blazor
  • WPF / WinUI / Avalonia
  • operator panels and offline work

Deployment

  • Linux / Docker / Nginx
  • on-premise / private cloud
  • CI/CD
  • testing, documentation and monitoring

How I work

How I work and what you can expect

You talk to the person building the application

No layer between a salesperson, a delivery manager and whoever eventually builds it. What is agreed in a call does not get lost in translation, because there is no translation step.

Documentation and repository on the client side

The repository can live in your organisation account from day one. The scope of the work handed over is set in the contract.

Straight answers about what not to build

If the problem can be solved with a spreadsheet, a configuration change or an off-the-shelf product at a fraction of the cost, you hear it during the free intro call — before any invoice exists.

Systems designed to be taken over

I assume somebody else will maintain this eventually. Hence the emphasis on readable architecture, tests and recorded decisions.

Experience across different environments

As a subcontractor, I have had the opportunity to build solutions for different clients and in different technologies. I will be happy to discuss them in person.

4
stages, each available separately
1–2 wks
cadence of working increments
handover
repository and documentation available to the client
EN / PL
project and documentation languages

Frequently asked

Before you write — the answer may already be here

An off-the-shelf system forces the company to bend its processes to fit the product. Custom software reverses that: it is built around the processes that already exist, integrates with what is already running, and carries no licence fees for features nobody uses. It pays off where the process itself is a competitive advantage, or where no packaged product covers it without expensive workarounds.

Yes. Each of the four stages is a self-contained service with its own deliverable. A common scenario is analytical consulting that ends with a requirements document and an estimate, on the basis of which the client decides whether and with whom to build the system. There is no obligation to continue.

Discovery and requirements analysis typically takes 2–4 weeks. A first production-ready release with a deliberately narrow scope usually lands within 2–4 months. A full enterprise-grade system with integrations and data migration runs 6–12 months depending on scope, delivered iteratively rather than in a single drop at the end.

The scope of economic copyright transfer is set in the contract. The repository, documentation and configuration can live in the client organisation from the start.

Yes. Engagements run remotely across Poland and internationally, in Polish or English. On-site workshops during discovery and acceptance are possible where the nature of the project calls for them.

The default is a full handover: repository, architecture documentation, runbooks and monitoring configuration, so that any other team can take over development. Independently of that, maintenance, feature development and support agreements with a defined response time are available.

Contact

Describe the problem — you get a straight answer on whether it is solvable

The first conversation is free and carries no commitment. A few sentences about what is not working today is enough.

office@mastras.com
+48 509 59 53 59
Remote — Poland, EU, international projects
Mastras · VAT ID PL5931285183
This field is required
Please enter a valid e-mail address
The more specific the situation, the more useful the first reply will be. Please describe briefly what this is about

This form sends nothing to any server — it opens your own mail client with the message pre-filled. The site uses no cookies and no analytics, and processes no personal data on its side.