Abdurahim Hudulov
AI · Product · Engineering
Background in AI evaluation, data quality, annotation and LLM reliability — exactly the hard part of compliance AI: grounding, hallucination control, confidence scoring and human-in-the-loop.
About
Compliance AI is not about text generation — it is about trust, evidence and review. That is exactly what our background is built for.
Origin
Large buyers automated sending their compliance questionnaires long ago. On the supplier side, a person still sits with Excel, Outlook and old answers. Software was built for the buyer — not for the supplier who has to answer. That is the gap Attestria closes.
Team
AI · Product · Engineering
Background in AI evaluation, data quality, annotation and LLM reliability — exactly the hard part of compliance AI: grounding, hallucination control, confidence scoring and human-in-the-loop.
Go-to-market · German market
Runs customer conversations and outreach to German Mittelstand companies — in German, in the region where our first customers sit. The Baden-Württemberg industrial belt.
Principles
Format detection
We recognise the format — TISAX, EcoVadis, LkSG or a custom Excel — and split it into individual questions.
Retrieval + grounding
We search your documents and build every answer from what is written there — with a reference to the exact paragraph.
Human-in-the-loop
Confidence scoring flags the uncertain. Risky answers always go to a person before anything leaves.
Location
The suppliers we want to serve sit right here. We are in Heilbronn, in the heart of the Baden-Württemberg industrial belt.
Honestly
We are early. The thesis is strong, the details are working hypotheses. We would rather name them openly.
Next step
Let’s talk for 20 minutes. We listen to how compliance runs at your company today — and whether Attestria can take over the repetitive part.