Founder-led studio · Maritime & process automation

AI systems for operations where mistakes are expensive.

Z-Deck turns a shipowner's Excel-and-WhatsApp office into one system: certificates, laytime, noon reports and voyage economics. Built by a chief mate who spent ten years on the bridge before writing the software that replaces the clipboard.

Python · PydanticAI · FastAPI Docker · n8n · Playwright Based in Türkiye → Bulgaria
What Z-Deck builds

Three problems, solved properly.

Not dashboards for their own sake. Each of these removes a specific, measurable piece of manual work from an operational day.

01

Certificates that never lapse

Crew documents and ship certificates checked against the flag's own rules. Lead times are derived per flag, so a Panama endorsement that takes 45 days is flagged 75 days out. Close, snooze or renew from one screen.

Flag rulesTR · PA · MTNightly reconciliation
02

Laytime you can defend

Statement of facts typed in seconds, GENCON laytime computed with the charter party's own terms, and a PDF or Excel statement for the charterer. Inputs and rules are stored with the result, so the same answer comes back in a dispute six months later.

GENCONSHINC / SHEX / WWDPDF statement
03

Voyage economics from the master's email

Noon reports reconciled against fuel remaining on board, agents' disbursement accounts tracked proforma to final, both feeding the voyage result. Extraction pre-fills the form; a person confirms before anything is written.

Noon reportsDisbursement accountsVoyage P&L
Selected systems

Things that actually run.

Every item below is either a live deployment or a public repository you can read. Nothing here is a mockup.

Live

MarOps — Shipowner Back Office

Multi-tenant back office for small dry-bulk owners: certificate alerts on each flag's own rules, evidence-grade GENCON laytime with PDF statements, noon-report fuel reconciliation and voyage P&L. Route planning now draws the live AIS track, with port cards from the World Port Index and navigational warnings along the route. 231 automated tests, in production.

Django 5PostgreSQLHTMX
In use · private

Passage Plan — Bridge Decision Support

A personal tool that turns a GPX route into a full passage-plan draft: VTS areas and reporting points as real polygons, live NAVTEX, Open-Meteo weather, EMODnet depth and TSS checks, zone-based UKC with squat and CATZOC, MARPOL ECA fuel changeovers, a master-approved A.893(21) checklist, RTZ/GPX and a PDF with map. Advisory; not a replacement for the ENC or the company SMS. 1,026 Python and 313 web tests.

FastAPIReactMapLibre
Live demo

Fleet Simulation

Three synthetic general-cargo ships across thirty-one days of hourly positions, noon reports and port SOFs — the exact shape of data a shipowner's office lives on. A map with a day slider, fuel and speed charts, and the voyage result MarOps computes from it. Runs in the browser, English or Turkish.

MaritimeLeafletNoon reports
In development

Noon Report System

An n8n and PydanticAI pipeline that turns free-text maritime noon reports into structured records with PDF and Excel output. The problem I lived with for ten years, finally automated.

n8nPydanticAIMaritime
Private repository
How to work together

Three ways in.

Step 01

Discovery sprint

One or two weeks. I map the actual workflow, find where the hours disappear, and come back with a scoped automation plan and an honest estimate — including the parts that are not worth automating.

Step 02

Build and pilot

A working system against your real documents and real edge cases, running alongside the current process until it demonstrably beats it. Typed, tested, containerised and handed over with documentation.

Step 03

Operate or hand over

Either I keep it running and extend it, or your team takes it over with the CI, the tests and the runbook already in place. No lock-in built on obscurity.

The founder

Ten years on the bridge, now writing the software.

I'm Oktay — an oceangoing maritime officer, chief mate, with ten years of watchkeeping behind me. I know exactly what a noon report costs in attention at 11:40 in the morning, because I filed them.

  • Ten years

    At sea

    Oceangoing watchkeeping officer through to chief mate. Cargo operations, stability, port state control, ISM audits, and the paperwork behind all of it.

  • The turn

    Python and data science

    Started automating my own reporting, then went properly into data engineering, statistical modelling and test automation.

  • Now

    Z-Deck

    A founder-led studio for GenAI process automation, with maritime operations as the vertical I understand best.

Contact

Tell me what's eating the hours.

Describe the process that keeps costing you a day a week. If I can automate it I'll tell you how — and if I can't, I'll tell you that instead.

Reply within 24 hours English or Turkish