Data scientist and project manager. Six years building real machine learning systems. By day at Fraunhofer ISI. Independently, I advise European teams on AI initiatives — reviews, EU AI Act readiness, and workshops.
A short selection. I share outcomes where I am allowed to.
A multi-agent platform that helps small and mid-sized companies see what is coming next. Agents scan signals, weigh evidence, and write decision-ready briefings.
A tool that spots emerging technologies before they hit the trade press. Bibliometric mining and language models cluster weak signals into clear trend cards.
Used by innovation teams across the EU public-sector network.
Demand forecasting for a German logistics company.
12% better than the ARIMA baseline.
A two-week diagnostic on a retrieval pipeline that was giving wrong answers in production. New re-ranker, new chunking strategy, proper evaluation set.
Citation accuracy 0.41 → 0.78. Wrong answers down by ~70%.
Risk classification, technical documentation, and human-oversight design for a high-risk health-tech system.
Passed internal compliance review on the first try.
Time series, forecasting, and NLP for strategic foresight.
Independent advisory for European teams shipping AI. Async-friendly. English and basic German.
Architecture reviews, system audits, and second opinions on AI initiatives. I review what your team builds — what works, what's risky, what to change first.
Compliance assessments and gap analysis against the EU AI Act. Risk classification of your AI systems and a prioritized remediation roadmap aligned with current EU guidance.
Hands-on training and ongoing advisory for technical teams. Production AI practices, evaluation methods, and AI governance — built around your context.
Something else in mind? Just ask. Every engagement is scoped individually.
Five steps. No proposal templates, no agency layers.
Thirty minutes. What you are building and what is stuck.
Within 48 hours. Clear deliverable, fixed price.
Shared tool, a Slack channel, weekly written update.
Code, evaluation, docs. Demo at every milestone.
Training for your team and two weeks of post-launch fixes.
A small selection. References available on a call.
In the first week, Bhavesh found three failure modes in our RAG pipeline that we had been chasing for months. The evaluation set he built is now the standard across our other ML services.
We needed someone who could actually read the EU AI Act and turn it into a build plan. Bhavesh did both, calmly, in three weeks. He saved us six months of legal back-and-forth.
A rare mix: senior judgment, a real publication record, and someone who actually writes the code. He pushed back where our internal lead would not have. That is why we kept him on retainer.
A thirty-minute intro call. No pitch. If I am not the right person, I will point you to someone who is.