Applied AI
that ships.

BI stack that replaced a €50K/yr controlling role.

A German marketing agency with no end-to-end view of billable efficiency. Over ten weeks I built 8 Power BI dashboards, 23 KPIs, and a Node middleware bridging Absence.io and Toggl. First month of deployment moved team efficiency from 49% to 65%.

• +8 Power BI dashboards, 23 KPIs — project + service teams
• +Node/Express middleware for HAWK + API-key auth bridging
• +49% → 65% project-team efficiency (Feb → Mar '26)
• +186-hr PoC converted to paid contract, now in phase 2

"It took me three years and it took you three weeks."

Meeting notes to sprint backlog in five minutes.

Planning meetings generate messy transcripts. Jira expects clean epics, stories, acceptance criteria, and estimates. This tool bridges the two — parses raw notes, drafts the ticket tree, files it against the Atlassian API. Built for teams tired of the Monday-morning ticketing tax.

• +Epic + story decomposition from meeting notes
• +Acceptance criteria drafting
• +Atlassian REST integration
• +Claude-backed extraction pipeline

The pipeline behind STOXX 600 climate-disclosure research.

A TRR 266 working paper needed financial-statement paragraphs from six years of STOXX Europe 600 annual reports, machine-readable. I built the scraper: PDF ingestion, LLM-backed section extraction, structured output. The paper cites the work in its acknowledgements.

• +STOXX Europe 600 annual + audit report coverage
• +PDF ingestion and section extraction
• +LLM-backed structured output
• +Cited: SSRN 4763140 (Müller, Ormazabal, Sellhorn, Wagner)

"We acknowledge excellent research assistance by Muhammad Bilal…"