murilo@cork:~$ cat profile.json
{
"role": "Data & AI Specialist · Automation",
"based": "Cork, Ireland",
"day_job": "Apple (contract) — CEMEA Sales BPR & Systems",
"freelance": "data & automation builds, open for work",
"languages": ["pt-BR", "en", "es", "it", "fr"],
"focus": ["ml in production", "automation", "analytics engineering"]
}
murilo@cork:~$ ./what_i_actually_do --verbose
[ ✓ ] AI & ML .............. churn scoring · forecasting · anomaly detection · NLP triage
[ ✓ ] AUTOMATION ........... month-end close · reporting pipelines · RPA on the edges
[ ✓ ] ANALYTICS ENG. ....... one certified source of truth, modelled properly
[ ✓ ] STORYTELLING ......... the recommendation made obvious, not just the numbers
[ ! ] NOTEBOOKS ONLY ....... not a deliverable. it ships or it does not count.The part of data work most projects skip: getting a model or an automation past the notebook and into somebody's Monday morning. Calibrated, backtested, idempotent, and handed over so the client owns it.
|
|
|
Full write-ups — the problem, the approach, and what actually moved — live at muriloreisz.com/work.
MuriloReisz.github.io is the one to look at first: the portfolio itself, Astro with zero UI framework, carrying a procedural ~730k-point WebGL cold open generated at runtime rather than downloaded as a mesh — a hooded figure made of code at a desk, and you fly through his monitor into the page.
Delivered work — the engagements behind the numbers:
Worked examples — anonymised composite scenarios rather than named client engagements, written up to show the approach and the reasoning. The figures are illustrative, and the site labels each one as such:
murilo@cork:~$ tail -f now.log
[building] agentic automations — LLM tool-use loops that do real back-office work
[learning] data storytelling · structuring an analysis so the call is obvious
[running] Cork AI Meetup — free, monthly, for local businesses
[open to] freelance data & automation builds