The work, in detail.
From healthcare and B2B platforms to live data and AI workflows. Each entry explains what the system does, how it works, and what it enables for the people using it. Client names are omitted where confidentiality applies.
Project list
Medical report digitization and analysis
A pipeline built to digitize, de-identify, and enrich medical reports. It is designed for a 600,000-document archive and an annual throughput of 200,000 reports. Interactive dashboards and a natural-language AI assistant complete the workflow.
Data extraction, validation, and AI analysis run in one workflow, with reports that need attention routed to staff for review.
Video-first recruiting, not CV-first
Candidates record a short video instead of submitting a CV. An AI pipeline extracts the desired role, skills, and search area, which the matching system uses to suggest nearby openings and profiles. One engineer covers product design, frontend, backend, and DevOps.
Azure AI processing configured in European regions, with payments and subscriptions built into the platform.
agent pipelines · co-founded companyFootprint: company reports built through AI workflows
A platform for configuring multi-agent workflows through a visual interface. Each agent has a dedicated model and MCP tools. Independent tasks run in parallel, then their outputs are assembled, critically reviewed, and translated when needed.
Workflows can be configured through the interface or the API without changing application code.
real-time & scaleLive sports data for a publishing group
The system ingests an external FTP feed, retains the source files, and parses them on a schedule into a normalized database. An API exposes the latest data to a library of customizable widgets embedded across the group’s publications.
One backend serves multiple publications, while each portal presents the same data in its own visual language.
The workflow, architecture decisions, and operational trade-offs behind each system.