PyData Eindhoven 2024
2024-07-11
Eindhoven, Netherlands
[PyData Eindhoven 2019, PyData Eindhoven 2020, PyData Eindhoven 2021, PyData Eindhoven 2022, PyData Eindhoven 2023, PyData Eindhoven 2024, PyData Eindhoven 2025]
who it's for
data scientists, machine learning engineers, sports data analysts, mlops engineers, data engineers, research engineers, product analysts
topics
retrieval augmented generation; LangChain workflows; vector search with Milvus; hybrid BM25 TF-IDF search; evaluating large language model frameworks; running quantized GenAI on desktop; scikit-learn pipeline caching; causal forecasting with DoubleML; BERTopic topic modeling; explainable AI with LIME; YOLO for sports analysis; switchback experimentation design
| title | author | length |
|---|---|---|
| The Levels of RAG 🦜 | Jeroen Overschie | 35m |
Builds a toy RAG over the conference schedule using LangChain, LangSmith and a vector DB, shows chunking and recursive splitters, adds hybrid search with… | ||
| Aftermovie | julak Pon | 02m |
A short event aftermovie that documents the seventh edition at the Philips Stadium and claims about 200 attendees and a successful cohosted program with julak… | ||
| Risks and Mitigations for a Safe and Responsible AI | Maria Medina | 35m |
Concrete claims and concrete artifacts matter: a risk framework that literally says Map Measure Manage, checklists tied to the EU AI Act, and measurable… | ||
| Scikit-Learn can do THAT?! | Vincent D Warmerdam | 35m |
Shows concrete capabilities and results: scikit-learn implements pipeline caching that can halve grid‑search time on a single machine, supports sparse inputs… | ||
| Enhancing Event Analysis at Scale: Leveraging Tracking Data in Sports | Yannis Moudere | 41m |
Pipeline ingests 250 MB of tracking per match with about 3.5 million coordinates and processes each game in 150 seconds per game using Python workers fed by a… | ||
| Predicting the Spring Classics of cycling with my first neural network | Rob Claessens | 34m |
A three‑layer neural network implemented in PyTorch trains on about 2200 samples (16 Classics, 5 years) to predict a continuous rank and reaches roughly… | ||
| Maximizing marketplace experimentation: switchback design for small samples | N. Alagoz, A. Mahfoudhi | 34m |
Conventional AB testing fails when units are few and effects are tiny so switchback design alternates treatment periods to cut interference and autocorrelation… | ||
| How I lost 1000€ betting on CS:GO with machine learning and Python | Pedro Tabacof | 32m |
Claims to have an exploitable edge in CS:GO betting, builds a Python web scraper collecting ~30,000 matches and three months of odds, engineers hundreds of… | ||
| Cloud? No Thanks! I’m Gonna Run GenAI on My AI PC | Dmitriy Pastushenkov, Adrian Boguszewski | 30m |
Claims a single AI PC runs GenAI in seconds by combining CPU GPU and NPU and using the OpenVINO Toolkit to optimize models and reduce dependencies. Demos are… | ||
| Causal Forecasting: How to disentangle causal effects, while controlling for ... | Marc Nientker | 32m |
Claims a deployable method that marries ML and econometrics to handle unobserved confounders by estimating causal elasticities, removing historical price… | ||
| Explainable AI in the LIME-light | Sanne van den Bogaart | 33m |
LIME claims to explain individual predictions by creating perturbations around one instance and fitting a local linear surrogate so you see which features… | ||
| BERTopic to Accelerate Ukrainian Aid by the Red Cross | Nina van Diermen | 32m |
BERTopic is applied to Telegram posts from Ukrainian refugees to turn messy messages into actionable labels, cutting manual labeling from 20 hours vs 2 minutes… | ||
| Evaluating LLM Frameworks | Ennia Suijkerbuijk | 34m |
An engineer from CM.com describes a Python evaluation framework for comparing LLMs and prompts for a production Retrieval-Augmented Generation (RAG) chatbot… | ||
| CV at the Dutch Tennis Federation: Utilizing YOLO to create insights for coaches | M.Brouwer, M.Kempe | 31m |
A one camera pipeline using YOLOv8 plus homography and U‑transform/UNet court detection produces 10 Hz player coordinates and rule based rally starts so a 5… | ||
| 🌳 The taller the tree, the harder the fall. Determining tree height from space.. | Ferdinand Schenck | 30m |
Claims a deep‑learning stereo pipeline implemented in PyTorch produces canopy height maps from sub‑meter stereo satellite pairs that substantially outperform… | ||
| Sonic Pi - Live Coding as a tool for next-gen education. | Sam Aaron | 01h 11m |
Sonic Pi is a free code-based music tool that targets classrooms and pro musicians, teaching sequencing, functions, loops and data structures while also… | ||
| nothing matches | ||