MARCO ROSSI
Zürich, Switzerland | marco.rossi@example.test | Swiss citizen

SUMMARY
Machine learning engineer, nine years, six of them running LLM-backed systems
in production for paying customers.

EXPERIENCE

Staff Engineer, Helvetica Data AG, Zürich — 2021 to present
  Built and ran a document-understanding pipeline processing ~40k insurance
  claims a month. Owned it end to end: retrieval, evaluation harness, cost
  control, and the pager. Took the on-call rotation for three years.
  An incident in 2023 was mine: a prompt change cut extraction accuracy and we
  did not notice for eleven days because the eval suite only measured latency.
  I rebuilt the harness around output quality and added a canary set. That
  class of failure has not recurred.
  Cut inference spend 60% by routing simple documents to a smaller model.

Senior Engineer, Nordwind Logistics, Basel — 2017 to 2021
  Route optimisation services in Python and Go. Moved deployment from hand-run
  scripts to CI with staged rollout and automatic rollback.

Software Engineer, Alpine Retail Systems, Bern — 2015 to 2017
  Inventory forecasting. First production exposure.

SKILLS
Python, Go, PyTorch, retrieval systems, evaluation harnesses, GCP, Terraform,
observability. German (native), English (fluent), Italian (fluent).

EDUCATION
MSc Computer Science, invented university, 2015.

AVAILABILITY
Three months' notice.
