B2B marketplace
EUDesigned and delivered backend services for a B2B marketplace, covering architecture, integrations, transactional workflows, and production operation.
BackendDistributed systemsAWSPostgreSQLIntegrations
Building backend and distributed systems that have to work.
I’m Ivan, a software engineer working across architecture, implementation, and production.
I build systems around distributed services, cloud infrastructure, data platforms, search, and applied AI.
Based in Spain, working remotely with international product teams.
- Backend
- Distributed systems
- Cloud infrastructure
- Data and search
- Applied AI
- Product engineering
Currently
Building backend systems for an AI-powered trade partnership platform for real-world industries.
Working across system design, backend implementation, integrations, data flows, and production reliability.
Engineering principles
- Understand constraints before choosing architecture
- Prefer simple designs that hold up
- Make system behavior visible
- Test behavior, not only mocks
- Automate boring parts; document sharp edges
Selected Work
Some systems I’ve built, fixed, or helped scale. Client details are intentionally limited.
Publishing and AdTech platform
USABuilt and evolved data pipelines and backend services for publishing and advertising workflows, including third-party integrations, search, reporting, and operational automation.
Data pipelinesSearchIntegrationsCloudBackend
Legal document platform
UKWorked on backend systems for document ingestion, conversion, metadata management, search, and real-time collaborative editing.
Document processingElasticsearchReal-time systemsBackendData
Notes
Short notes on backend systems, architecture, applied AI, and whatever technical rabbit hole I’m in.
Shared reading
Agent swarms and the economics of model roles
Cursor’s experiment separates planning from execution: stronger models handle decomposition and difficult decisions, while faster models do most of the implementation. The interesting part is not only parallelism—it is keeping each agent’s context narrow and purposeful.
- AI agents
- Agent swarms
- Model economics