consulting
systems that actually run in production
i work throughlaava (opens in new tab) as a hands-on technical partner — building ai agents, integrations and cloud infrastructure that run reliably in production.
not ai dev, not chatbots, not strategy decks.
what i build:
context — find the right information, respect permissions, cite the source.
reasoning — classify, compare, check, prioritize, with confidence scoring and escalation.
action — connect to the real workflow, with audit trail and least-privilege access.
all on eu cloud: kubernetes, ci/cd, monitoring and governance.
how i work:
audit — fixed-price discovery of 2–3 real ai/infra opportunities. from €1.5k (credited to your build).
build sprint — a production system delivered end to end: infra, agents, integrations, monitoring, handover. from €10k.
operate & sla — i own it and keep it running, with a response guarantee. from €1.5k/mo.
proof:
energy-flexibility scale-up
friction — fast-growing, but hosting didn't scale, data was siloed, and the back-office ran on manual, people-dependent work.
built — kubernetes migration on eu cloud (iac, ci/cd, monitoring), an ai gateway, receivables + helpdesk agents, and a data integration layer.
result —
manual receivables prep: hours → minutes per cycle
one scalable eu platform, replacing ad-hoc hosting
multiple agents live in production — not demos
maritime manufacturer
friction — a mountain of technical specs and drawings; document review was slow, expert-heavy, and error-prone under deadline.
built — sovereign ai infrastructure, ai ocr pipelines for document ingestion, and permission-aware spec/document-review workflows, run on-cluster under a formal sla.
result —
document review: 2h → 15min per item
runs on sovereign infrastructure — data stays in-house
every step traceable, under a defined sla owned by laava
who it's for:
best fit — ~20–300 people, real revenue, messy operations, and a founder/cto/coo who can move.
not a fit — "we just want a chatbot", pre-revenue, or demo-ware.