PyData Eindhoven 2023
2023-11-30..2023-12-01
Eindhoven, Netherlands
[PyData Eindhoven 2019, PyData Eindhoven 2020, PyData Eindhoven 2021, PyData Eindhoven 2022, PyData Eindhoven 2023, PyData Eindhoven 2024, PyData Eindhoven 2025]
| title | author | length |
|---|---|---|
| Aftermovie | 02m | |
A short aftermovie that lists concrete conference facts and community details rather than big promises: two keynotes, 25 speakers, 40 volunteers, more than 90… | ||
| Leveraging Conformal Prediction for Calibrated Probabilistic Time Series Forecasts | Inge van den Ende | 32m |
Claims a procedure that gives valid finite‑sample coverage for probabilistic energy price or production forecasts under exchangeability, using a held‑out… | ||
| Reactive Notebooks for Python | Stephan Sahm | 34m |
An interactive notebook platform that actually re-executes dependent cells when inputs change and resolves execution order reactive cells auto-rerun… | ||
| Keynote: Solara simplifies building complex dashboards. | Maarten Breddels | 44m |
Solara claims to let teams build complex data apps in 100% Python by combining a ReactJS‑style component API with reactive variables auto-trigger rerenders and… | ||
| 5 ways to fail with time series | Alyona Galyeva, Dr. Sebastian Werner | 34m |
Five concrete pitfalls and fixes: mismatched sampling frequency matters creates artifacts when resampling; naive train/test splits leak future information so… | ||
| Python in the browser | Chris Laffra | 33m |
Demo of PyScript shows runnable Python apps in the browser: a tiny UI toolkit ltk (~600 lines) plus three apps (microlog, a spreadsheet that is “Excel in… | ||
| Taming the Machine: Basics of ML Models Training and Inference Optimization | Vladimir Osin | 32m |
Covers concrete knobs and measured wins: mixed precision (≈3× speedup, sometimes 5×), batch-size + optimizer tuning, gradient accumulation and checkpointing to… | ||
| Your best Bet: Effortless MLOps with Python Models in | Dorian Van den Heede | 33m |
Claims that dbt 1.3 plus Python models run on Databricks can implement a full ML pipeline: data prep in SQL, feature snapshots, experiments (cross‑validation… | ||
| Structuring automated visual inspections with Django and gRPC | Thomas Woudsma | 33m |
An MLOps prototype uses Django admin + models to define Product → InspectionView → InspectionTask (x,y,width,height) plus Label records, validates rectangles… | ||
| Team Red: Breaking Large Language Models | Mickey Beurskens | 30m |
Claims that prompt injection enables real data leaks and control of LLM-driven workflows, demonstrated by URL exfiltration when browsing is enabled and by… | ||
| Building & Deploying LLM Apps | Max Pumperla | 36m |
Claim: Llama 2 handles text summarization at cost roughly 30× cheaper than GPT‑4 while matching GPT‑3.5 Turbo on many tasks. Tools: Ray is used to parallelize… | ||
| Building Fast Packages Faster: Julia as a Backend to Python and R | Joris K. | 34m |
Concrete claim: Python and R wrappers diffeqpy and diffeqr bridge to DifferentialEquations.jl from SciML, install Julia and GPU stacks automatically, and run… | ||
| A New Way to Manage Python Environments | Ruben Arts | 32m |
pixi claims a single binary that bridges conda and pip, auto-generates reproducible lock-files and runs on Windows Mac and Linux; it uses a rust… | ||
| and time zones: everything you need to know | Marco Edward Gorelli | 30m |
Polars gives concrete answers: it parses CSVs with parse_dates=True, converts offset-aware strings to UTC, and requires an explicit timezone dtype when local… | ||
| Enhancing Quality Control Efficiency: A Dynamic Risk Threshold Approach | Erdem B., Bouke H. | 29m |
Daily inflow of about 15,000 items/day meets a QC capacity of 200 so a system combines a per-item risk score (daily-retrained Random Forest with SHAP… | ||
| Background removal without background knowledge | Ariadna Kramkovska | 28m |
Builds an actual pipeline: generates training pairs by compositing programmatic data generator from PNGs and hand-labels 25 examples to prototype a masker… | ||
| Active Teaching, Human Learning | Vincent D Warmerdam | 37m |
Empirical claim: active learning often helps but sometimes hurts depending on number of classes, tiny random starts, or huge batch sizes; synthetic… | ||
| Beyond the Continuum: The Importance of Quantization in Deep Learning | Adrian Boguszewski | 34m |
Quantization shrinks models and speeds inference with concrete gains: weight compression cuts LLaMA‑7B from 25 GB to 6 GB and Dolly from 44 to 11 GB… | ||
| Adjusting 2D prediction models to be usable for 3D objects | Vera van der Lelij | 35m |
Claims a workable pipeline: starts from STL point‑clouds, uses open3d for centering and ICP + fast global registration to align scans, learns… | ||
| How to build production-ready data science pipelines with | Ionut Barbu, Tim Brakenhoff | 33m |
Kedro structures code with a project template, YAML data catalog and node/pipeline functions, and integrates with MLflow for model artifacts and versioning. A… | ||
| Reinforcement Learning for Food Waste Reduction within Albert Heijn | Patrick de Oude | 32m |
A deployed system uses a discrete state space and offline Q-learning to set three staged markdowns 25/40/70 markdowns for items near expiry and learns from… | ||
| Dataset enrichment using LLM's ✨ | Jeroen Overschie | 34m |
A pipeline extracts hidden housing features from listing text into JSON/Pydantic objects using openai and instructor, compares GPT-3.5 GPT-4 and PaLM on ~100… | ||
| Keynote on Polars Plugins | Ritchie Vink | 49m |
Polars is a Rust query engine that uses a persistent thread pool, work stealing and cache‑friendly Arrow buffers to do lazy, parallel execution and query… | ||
| Reactive Notebooks for Python | Stephan Sahm | 36m |
Reactive Python notebooks implement automatic dependency tracking so cells re-run only when needed and claim impossible to reach an invalid state by design… | ||
| Bart Steverink & Laurens Schinkelshoek | Bart Steverink, Laurens Schinkelshoek | 29m |
Healtphlus.ai builds PERISCOPE©, a clinical decision support that predicts post‑surgical infection probability at 7 and 30 days using hospital EHR structured… | ||
| SHAPtivating Insights: unravelling blackbox AI models | Maximilian M. | 35m |
SHAP analysis on a tabular bank transaction classifier (XGBoost) exposes concrete findings: a single feature creates 100% accuracy due to label leakage, that… | ||
| Brain-Inspired Natural AI: Unlocking Intelligence | Bart van Erp | 25m |
Claims a concrete route from the Free Energy Principle to low‑power, online‑learning agents and supplies an actual toolbox RXInfer (Julia) that automates… | ||
| nothing matches | ||