Available full-time from October 2026 · Germany / EU remote
Three systems in production, one method in daily industrial use.
Autonomous LLM pipelines, a self-hosted agent stack, and a sub-second knowledge base - each running today, each with the numbers to back it. Here is what the problem was, what I built, and what it does now.
48 seconds, three of the systems below - real terminal recordings, no mockups
Current role - step by step
What I do at Kyocera, in the order I do it.
AI and digital transformation for an industrial B2B organisation - the technical track of a corporate AI programme, run as four repeatable steps.
01Map
Process discovery through stakeholder interviews and workshops - quality, customer service, sales, change control - captured as BPMN process models.
02Find
Use-case identification and automation-gap analysis: findings documented with source evidence and structured into a catalog management can prioritise.
03Solve
Working automation, not slideware: an email-assistant agent for customer service with human-in-the-loop by design, plus M365 automation flows with telemetry and QA gates.
04Prove
Acceptance tests before anything ships, ROI business cases, and rollout documentation aligned with the EU AI Act and GDPR - plus contributions to the corporate AI usage policy.
This is the technical track of a company-wide AI programme - the same method I run in consulting engagements, applied every working day.
ProblemReading market and news feeds by hand does not scale and does not run at 3am.
BuiltAn autonomous pipeline that ingests market and news data, runs the LLM analysis, and publishes signals to Telegram on a schedule with zero manual steps. The model layer is vendor-neutral - Anthropic, OpenAI, DeepInfra, or a local Ollama model - and a shadow-evaluation gate only promotes a cheaper model once it matches the reference on live data.
deepsignal
docker ps - all services healthy on the server
~95%lower model cost after the shadow-eval gate
hardened by backtesting + a 54-agent adversarial review / self-hosted in Docker / months in production
ProblemOne assistant cannot be an expert at everything, and a cloud assistant does not run on your own hardware or your own terms.
BuiltA self-hosted assistant that runs 24/7 and coordinates a library of specialist agents through a meta-orchestrator, with custom MCP servers, voice in and out, and direct email and calendar tools.
"why isn't the graph loading on my phone?"
berezka-site-optimizer - owns berezka.me end to end
one orchestrator, 12 division directors, 200+ specialist agents - hover any node for its name, one real routed example traced in bronze
PythonClaude APIMCPWhisperDocker
🔎In production
Hybrid RAG knowledge base
ProblemA 4,500-note knowledge base is only useful if the right answer comes back in one second, not by scrolling through files.
BuiltA hybrid retrieval system - dense embeddings (bge-m3) plus sparse full-text search (FTS5) plus a reranker - reindexed nightly and served by an always-on daemon.
ProblemIndustrial companies know AI matters. The hard part is deciding where it actually pays off - before spending the budget.
BuiltAI and digital transformation work for an industrial B2B organisation: business-process analysis and structured, compliance-aware evaluation of AI use cases. The method behind it is public research - my KIT master's thesis builds a multi-criteria decision model for AI adoption in manufacturing.
M.Sc.KIT thesis: a multi-criteria decision model for AI adoption in manufacturing
process analysis / structured use-case evaluation / EU AI Act + GDPR aware
BPMNMCDM / AHPEU AI ActGDPRM365
The knowledge base behind it
4,540 notes. 13,024 connections. One graph.
The Obsidian vault that powers the hybrid RAG system above - every dot is a note, every line is a link between two of them. Titles hidden on purpose; the structure is the point. Top 700 nodes shown.
4,540 notes13,024 links~2,000 indexed for retrievalreindexed nightly
The graph didn't load - probably a network hiccup.
Let's talk
Available full-time from October 2026.
Based in Germany. If you need someone who can take an idea to a running, self-hosted system - not a slide deck - let's talk.