01 Consulting and leadership
- Design the architecture before anyone writes code
- Audit a codebase and run technical due diligence
- Step in as tech lead or team lead
- Break a feature into tickets a developer can pick up without asking twice
- Run refinement and estimates with the team
- Track priorities, dependencies and whoever is blocked
- Estimate a project before it starts, and say what is not in it
- Present the design and the sprint review to the client and their architect
- Write your interview tasks, run the interviews, then onboard whoever you hire
- Set up CI, code review and a release flow the team can follow
- Plan how a team starts working with AI
02 Software engineering
- Build and run backend services in Java and Spring Boot
- Run a business process as a BPMN workflow instead of hiding it in code
- Make the build enforce the architecture rules, so review does not have to
- Test against real databases and queues with Testcontainers, not mocks
- Scan every build for vulnerabilities and code quality with Aikido or SonarQube
- Move business logic out of PL/SQL into services, one piece at a time
03 Data and integrations
- Integrate services over Kafka, BPMN, REST and webhooks
- Generate the API contract from one OpenAPI spec, so backend and frontend cannot drift
- Build the master data system that the rest of the company reads from
- Put one API over data that arrives in a different shape from every supplier
- Model the data, write the migrations and build the reports
- Connect payments and other third-party APIs
- Push live data to the browser over Kafka and server-sent events
04 AI implementation
- Ship LLM features that extract, classify, search and generate
- Automate an existing process with agent workflows
- Build an MCP server so an assistant can reach your data and your tools
- Put a retrieval layer over your documents, so answers cite your own content
- Write and version the prompts, and keep them working when the model changes
- Cut cost with model routing, or run local models when data cannot leave the building
- Measure prompts against real data instead of a demo
- Teach the team Claude Code and Codex
05 Web delivery and SEO
- Build marketing sites on Astro and Cloudflare Pages
- Build a landing page around one action
- Set up and run a WooCommerce shop
- Write the copy for pages and products
- Fix technical SEO: crawlable structure, honest metadata, fast on a phone
- Do AI SEO: structured data and copy an assistant can quote back
- Publish an llms.txt, so an assistant reads the site the way you wrote it
- Let the AI crawlers in on purpose, and say which ones
- Track what search and AI answers actually send you
06 Infrastructure and operations
- Host the site and run the domains
- Set up CI/CD, containers and infrastructure as code, with a rollback that works
- Deploy to AWS, Azure, Scaleway or Hetzner
- Work with Kubernetes and Terraform
- Administer Linux and macOS machines, backup and monitoring included
- Watch it with logs, metrics and alerts, so you do not learn about an outage from a customer
07 Marketing and growth operations
- Set up a CRM and automate the lead flow
- Send email campaigns and keep them out of spam
- Run Meta Ads, from creative to reporting
- Plan and publish organic social, graphics and copy
- Build dashboards that show revenue, not likes
08 Video and visual content
- Render short video and reels from code, in batches
- Shoot and edit drone footage
- Produce social graphics as a template system, not one-off files
09 Selected work
- Onboarding platform for a retail bank, Denmark. Greenfield build: customer onboarding plus an advisor portal for bank staff. BPMN workflows coordinate Spring Boot microservices behind an API-first contract. Integrations cover national eID, customer data, PEP and sanctions checks, credit assessment. I co-architected the service boundaries and stay hands-on with workflow execution and dynamic step rendering. I lead the developer side of a cross-functional team: refinements, task breakdown, PR review, technical interviews. CI/CD on Azure DevOps, infrastructure on Terraform.
- Trading platform, capital markets, Poland. Live Warsaw Stock Exchange data goes over Kafka and server-sent events to the trading apps, desktop as well as web, while the platform migrates off the legacy exchange protocol. Reactive services on Spring WebFlux, Oracle underneath, OAuth2 and JWT on the edge. Kafka carries events between services, instances and domains. I also put the quality gates in: SonarQube, Testcontainers around the integration tests, ArchUnit guarding the architecture. Liquibase owns the schema, GitLab CI/CD ships into Kubernetes, Prometheus and Grafana cover what happens after.
- Clearing and settlement, the Netherlands. Central counterparty platform for securities financing. First cleared model for European SFTs, covering equities and ETFs across 19 markets. I built its master data system from scratch, the single source of truth for participant and venue data. Around it: OAuth2 APIs for cross-domain exchange, Kafka events, a full audit trail, a back-office GUI. PostgreSQL under it, Keycloak issuing the tokens, OpenTelemetry into Prometheus and Grafana. Infrastructure on Terraform and Kubernetes.
- Core banking, Poland. Three years of backend work. Java and Spring Boot services on a hexagonal architecture. Wide banking REST APIs, integrations over JMS queues and webhooks, a Vaadin back-office application. Part of it was taking logic out of a PL/SQL monolith, the rest was building what replaced it. I owned the collateral module, a separate application for loan and credit line collateral, and led the developers in that domain. I also improved the developer sandbox that API partners test against. Oracle and PostgreSQL underneath, Flyway on the schema, WSO2 and WebLogic JMS at the edges.
- Vehicle data platform, automotive, United Kingdom. Universal API that normalised Renault and Mercedes feeds into one format, so a consumer reads one contract instead of one per brand. High-throughput streaming underneath, running on Kubernetes with Prometheus and Grafana over it. I moved the key services from AWS to Azure: DynamoDB, SQS and S3 on one side, CosmosDB, Service Bus and Data Lake on the other. Monitoring got better and an error took less time to resolve. I prototyped the Kafka streaming for high-volume vehicle telemetry. Spring Batch, Spring WebFlux and OpenFeign in the services, Keycloak on the APIs, Flyway over PostgreSQL, Cucumber and WireMock in the tests.
- Education business, paid acquisition and AI, Poland. I run the paid side of a high-ticket online course. One launch of it brought in over 500,000 PLN net of signed contracts. Meta campaigns from prospecting through live-event signups to deadline retargeting. Hundreds of creative variants, the ones to launch picked by blind audit rather than taste. Underneath: server-side conversions, per-ad-set attribution, a webinar funnel with a qualifying survey, email automation. I built the CRM the business sells through and I run the delivery around it. I also run its AI side: workshops on Codex and coding agents for the team, and a one-human-many-agents harness where I brief agents, review what comes back and keep the work moving in parallel.
10 How I work
- B2B contract, VAT invoice.
- Specifications, plans and decisions written down before the code.
- Own hardware and tooling.
- Remote, from Poland. Polish and English.
- Priced per scope or per day. Fixed-scope work quoted upfront.
- I carry responsibility for the result of my work.
11 Contact
- Business
- ANDREASIK.COM Sławomir Andreasik
- Seat
- Rzeszów, Poland
- NIP
- 8133605545
- REGON
- 389011203