A selection of projects our team has designed, built and run end-to-end — for enterprise clients across travel, fintech, HR-tech and energy.
Problem: a large Java monolith made every new feature slow to ship and drove up server costs.
What we did: redesigned the architecture around Spring microservices in Docker and set the template for all future features — API gateway, config server with Git integration, OAuth2/JWT security, discovery/registry services, Hazelcast caching, Feign clients, ELK logging and Grafana dashboards, on a new CI/CD pipeline.
Result: a scalable, observable platform that the team could extend feature-by-feature without touching a monolith.
Problem: consolidate payments across a high-volume e-commerce business with strict reliability needs.
What we did: designed and implemented a new centralized payment system on a fully asynchronous microservices architecture, including Klarna Pay Later, coordinating across several teams.
Result: a resilient, decoupled payment backbone running in production on Google Cloud.
Problem: capture and structure the entire DACH job market in near real time.
What we did: built a Java crawler running ~16,000 threads across a 30-server farm, using deep-web navigation, form submission, JavaScript evaluation and feature extraction. Every job was parsed, enriched and cached.
Result: more than 2 million jobs updated daily and a peak of 200M+ jobs in the database, feeding a live job-search portal.
Problem: validate and classify millions of crawled pages, and extract structured detail from free text.
What we did: built a multi-layered ML pipeline that extracted real content from each page, classified it (HR / non-HR, job / non-job, then job type), and pulled out requirements, benefits and salary range. Exposed via REST API.
Result: clean, categorized, queryable job data powering search and matching at scale.
Problem: unify booking across many providers with payment processing built in.
What we did: built backend booking microservices integrating a multitude of service providers, including credit-card processing — plus an OCR API for extracting data from scanned receipts and invoices (with image pre-processing).
Result: a single API surface powering multi-modal trip booking and expense capture.
Problem: let users plan and book a journey across local buses, trains and flights in one flow.
What we did: built the front-end trip planner, with each segment configurable and bookable against the booking microservices via REST.
Result: a single, bookable itinerary spanning several modes of transport.
Problem: mitigate energy-market balance for multiple operators across Nordic countries and the US, under high security constraints.
What we did: built enterprise-level applications spanning complex service interactions and large-scale transactions, at staff-engineer level.
Result: reliable, secure market applications operating across multiple operators and regions.
Tell us what you're building and we'll walk you through the most relevant work in detail.