Meetup at Visma

An evening of talks at Visma: custom LLM benchmarks, survival models, and Kalman filters — plus snacks, drinks, networking, and the (in)famous PyData Helsinki quiz. Please be on time! Latecomers may be unable to get in, since Visma personnel need to be there to let you in through locked doors.

Sponsored by Visma.

Schedule

• 17:30 — Doors open
• 18:00 — PyData Helsinki Opening
• 18:05 — Welcoming words from the host
• 18:10 — Talk: Mistral LLM – Custom Benchmarks (Allan Nevala)
• 18:40 — Talk: What is a survival model? (Babäk Firoozi Fooladi)
• 19:10 — Break — snacks, drinks, networking
• 19:40 — The (in)famous PyData Helsinki quiz
• 19:50 — Talk: Implementing Kalman filter in Python (Anna Korolyuk)
• 20:30 — End of programme
• 21:00 — Continue discussions at a nearby bar

Programme

• Allan Nevala: Mistral LLM - Custom Benchmarks

  Allan shows his custom Mistral LLM benchmarks in Jupyter notebooks: the comparison methodology, the use of Pydantic AI, data visualisations, and how Mistral performed on various tasks relative to models from other providers such as Google and Anthropic. The results so far have been surprising.

• Babäk Firoozi Fooladi: What is a survival model?

  Ever heard of survival models, time-to-event models, or duration models? If not, join the meetup to get familiar.

• Anna Korolyuk: Implementing Kalman filter in Python - from linear to non linear

  Kalman filters power sensor fusion in drones, fintech, and robotics. This talk walks through implementing them in Python, from the linear case to non-linear variants.

Location

Map location

This event has taken place. The original event listing is on Meetup.