Enterprise finance consultant by day. I use hands-on building with AI to stay ahead of what I advise clients on - not as a developer, but as someone who needs to understand the tools well enough to direct them.
Day Job
SAP FI consultant helping large organisations navigate complex finance transformations. I sit at the intersection of process, compliance, and technology.
How I Think
I define the problem, design the solution, and direct AI to build it. The decisions are mine. I own architecture, testing, and deployment. AI handles execution speed.
How I Work
The discipline is the same whether the problem is a broken finance process or a missing data pipeline. Define it clearly, then build the right thing - not just the obvious thing.
Most solutions fail because they solve the wrong problem. Before anything else, I separate the stated problem from the actual one. Who is affected and how? What does a good outcome look like versus a merely complete one? What constraints are real versus assumed? A well-framed problem is often half-solved.
I define the architecture before touching any tool. What are the components? How do they connect? Where does data flow and where could it break? What edge cases will surface in production but not in demos? This is where most execution risk lives - and where thinking it through in advance pays off most.
I direct AI tools to implement the design, then test against real conditions rather than ideal ones. Does it handle bad input? Does it fail gracefully? Does it actually solve the problem for someone who did not build it? Iteration here is fast because the design is already clear - I am validating assumptions, not discovering them.
A working deployment is the only honest test. I move from local to live quickly and treat the first real use as the start of learning, not the end of building. What did users actually do? What did the system do unexpectedly? The feedback from a live product is worth more than any pre-launch analysis.
Projects
Every project here started with a real problem. One is actively used every day. The others are deliberate learning exercises - the fastest way to understand a technology is to build something that would break without it.
On-premise AI that retrieves and cites exact sections of law - no cloud, no hallucinations.
Standard AI assistants either refuse to answer legal questions or invent rules that do not exist. German immigration law spans multiple statutes with hundreds of cross-referenced sections - a single eligibility question can require reading five different parts of the law at once.
A system that runs entirely on local hardware - no data ever leaves the machine, full GDPR compliance by design. It retrieves the most relevant legal sections for any question and generates an answer grounded only in those sections, citing exactly where each point comes from.
Log expenses and track family goals - from any device, with just a message.
Logging a family expense means opening an app, finding the right category, and typing it manually. The friction is small enough to skip - until weeks of spending disappear. Goals suffer the same way: written down once, never revisited.
A Telegram bot that reads a message or a photo of a receipt and writes the expense directly into a shared Google Sheet in under a second. Goals work the same way - one natural-language message creates a tracked entry. Zero apps, zero manual effort.
What enterprise cash visibility could look like if it were built today, not fifteen years ago.
Enterprise finance teams spend more time gathering data than analysing it. Cash sits across multiple systems and geographies with no single view. The tools that exist were designed for a world before real-time data and AI.
A working prototype of an ERP-agnostic treasury layer that pulls data via API without touching the underlying system, uses AI to surface insights, and gives a finance team a live command centre instead of a weekly Excel report.
More Work
100% automated daily episodes. AI-generated, zero operating cost.
Podcast production requires daily manual effort. This pipeline generates, voices, and publishes episodes automatically — no intervention at any step.
DemoTransparent credit scoring engine. Every calculation visible and interactive.
Credit scores are black boxes. This makes every calculation visible and interactive — Basel III methodology, fully auditable by anyone.
LocalOffline photo library manager with local face recognition. No cloud.
Cloud photo exports strip all metadata — timestamps, GPS, and face tags vanish. This recovers everything locally using offline OpenCV models.
LiveGamified German learning with interactive grammar and memory games.
Language learning apps are passive. This uses interactive games and spaced repetition to make German genuinely engaging to learn.
LiveZero-install multiplayer party game with <50ms Firestore sync.
Party games require downloads, accounts, and setup. This runs instantly in any browser — share a link, pick up your phone, start playing.