BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.evolutio.pt//pycon-portugal-2026//speaker//3CJS9
 F
BEGIN:VEVENT
UID:pretalx-pycon-portugal-2026-ZJGACA@pretalx.evolutio.pt
DTSTART:20260903T123000Z
DTEND:20260903T130000Z
DESCRIPTION:A snapped fridge handle. A missing washing-machine knob. A brac
 ket that broke off your scooter. The replacement part costs three pounds a
 nd ships in two weeks\, or you throw the whole thing away.\n\nThis talk is
  about teaching Python to fix it instead.\n\nThis talk walks through an op
 en-source Python pipeline that takes a photo of a broken object and produc
 es a 3D-printable patch automatically. Seven specialised agents handle cal
 ibration\, segmentation\, measurement\, thickness estimation\, mesh genera
 tion\, validation\, and printing. Each agent runs multiple competing compu
 ter-vision strategies\, then uses an LLM to choose the most trustworthy re
 sult through a consensus protocol. The validated mesh streams straight to 
 a BambuLab printer over MQTT\, and a finished patch comes off the bed by t
 he time the talk ends.\n\nThe whole system is built on Python\, using fram
 eworks such as FastAPI\, OpenCV\, Shapely\, trimesh\, SAM\, Depth Anything
 \, and a Three.js front end\, plus an LLM acting as the decision engine in
 side every agent. You'll see how far Python's ecosystem reaches when the o
 utput has to obey physics\, and where the interesting failures live.
DTSTAMP:20260819T142708Z
LOCATION:Auditorium
SUMMARY:From Pixels to Plastic: Multi-Agent Python Pipelines for 3D-Printed
  Repair - Brain Aboze
URL:https://pretalx.evolutio.pt/pycon-portugal-2026/talk/ZJGACA/
END:VEVENT
END:VCALENDAR