PyData Tel Aviv 2024
2024-09-25..2024-09-26
Tel Aviv, Israel
[PyData Tel Aviv 2022, PyData Tel Aviv 2024, PyData Tel Aviv 2025]
| title | author | length |
|---|---|---|
| State-Space Models and Deep Learning: Is there a new revolution on our doorstep? | Mike Erlihson | 46m |
Claims that state-space models give linear state transitions enabling parallelizable training plus fast O(n) inference for very long contexts; derives from… | ||
| Processing Biggish Data with DuckDB and Python | Yoav Nordmann | 25m |
Claims that DuckDB lets Python replace Spark/Trino for "biggish" workloads by running an in-process SQL RDBMS that avoids pandas memory limits and pushes data… | ||
| Adding Your Own Data Apps to JupyterLab | Daniel Goldfarb | 27m |
Step-by-step demo shows how to extend JupyterLab using TypeScript, lumino widgets, npm/JPM and the jupyter lab packaging tools, with a runnable repo at… | ||
| Supercharging CTR (Click-Through Rate) with Plug-and-Play AI Capabilities | Ronny Ahituv | 37m |
Claims a runnable, fully AI‑driven CTR pipeline that uses off‑the‑shelf cloud tools to generate creatives (generative models + genetic algorithms), estimate… | ||
| The TL;DR of EDA | Mor Hananovitz | 24m |
Automated EDA on a 200-wine dataset with 13 chemical/physical features uses pandas and quick seaborn plots, a profiling report (ydata/pandas-profiling style)… | ||
| Causal inference with Causallib | Ehud Karavani | 26m |
Argues that prediction alone misleads—example: a clinical risk calculator shows predicted risk rising after treatment—and demonstrates concrete methods and a… | ||
| Georgia on my Mind: NLP Meets Social Network Analysis for Exploring New Domains | Yuval Feinstein | 36m |
Method combines social network analysis and NLP to turn a few English Wikipedia seed pages into usable maps of a domain using networkx spaCy wikipedia. The… | ||
| A Shallow Introduction to Self-Attention | Alon Oring | 26m |
Starts from perceptrons and RNN failure modes and shows a concrete implementation: embeddings get projected into Q K V projections then scored by scaled… | ||
| The Dangerous Data Anonymization - Keynote | Ran Bar-Zik | 32m |
Data anonymization often fails in practice: simple masking with asterisks is reversible by joining multiple masked datasets, Netflix/IMDb-style linkage is… | ||
| Empowering ML Developers | Roman Olshanskiy, David Katz | 21m |
Platform lets ML teams ship custom web analytics from plain Python code and removes the need for deep SQL or cloud expertise, claiming to cut dependence on… | ||
| 2D ARIMA, Capturing New Trends for Distant Time Horizons | Yonathan Guttel | 21m |
Two-Dimensional ARIMA is a concrete method that pivots cohort-by-age revenue into a matrix and applies an iterative column-wise ARIMA so each predicted month… | ||
| Using Row Groups for fast filtering of large parquet files | Uri Mogilevsky, Menachem Kluft | 30m |
Shows that tuning Parquet row groups and layout cuts IO and cloud cost: reading footer metadata with pyarrow’s ParquetFile and adding per‑group stats (min/max… | ||
| Times and Dates in Pandas | Reuven M. Lerner | 28m |
Pandas code converts CSV strings to real datetimes with pd.to_datetime and read_csv(parse_dates=...), shrinking NYC taxi timestamps from ~977 MB as strings to… | ||
| Unveiling the Journey of NLP | Ortal Ashkenazi | 23m |
Claims concrete wins and clear limits: Transformer (2017) delivers parallel attention that captures long dependencies, pretraining plus fine‑tuning lifts SQuAD… | ||
| BertTopic, From Free-Text feedbacks to Calls for Action | Moran Reznik | 24m |
Shows a concrete pipeline that converts free text into actionable topics using sentence-transformers embeddings, dimensionality reduction with UMAP + HDBSCAN… | ||
| ChatGPT Goes Beyond Its Knowledge Cut-Off | Yoel, Shuki | 28m |
Live-coding that scrapes TechCrunch articles from the past seven days, extracts paragraphs with BeautifulSoup and pandas, and reduces ~2,300 rows to a smaller… | ||
| Optimizing Data-Driven Decisions | Aviv Vromen | 17m |
Builds an aggregation engine that continuously computes daily stored aggregations so features like “transactions in last 6 months” read at most 180 rows… | ||
| Building a Reproducible RAG Pipeline | Isan Rivkin | 27m |
Step‑by‑step build of a reproducible RAG pipeline that uses LangChain + Ollama + Hugging Face Embeddings, an S3-backed vector store and metadata time‑weighting… | ||
| AI, SQL, and GraphQL Walk into a Fertility Clinic | Shirli Di-Castro Shashua | 27m |
Compares three concrete architectures—LLM+SQL, LLM+GraphQL, LLM+RestAPI—with runnable Python chains that generate queries, execute them, and validate results… | ||
| APL-Inspired Techniques for Advanced NumPy | Eran Krakovsky | 27m |
Shows concrete APL techniques applied to NumPy and other array libraries to eliminate Python loops, cut allocations, and speed up code: replaces nested… | ||
| SaaS Compliance Automation with a Python Stack | Tomer Doitshman | 21m |
Catalogs tens of thousands of SaaS apps, builds per-app security/compliance profiles from netflow, web search results and news, and reduces manual work to a… | ||
| Ibis framework, Making data science work at any scale | Omri Fima | 28m |
Shows that Ibis translates familiar dataframe code into SQL and scalable engines so the same notebook runs locally and on large backends with minimal changes… | ||
| Securing Language Models Against Prompt Injection | Michael Ethan Levinger | 24m |
Lists three concrete attack classes: prompt injection (examples include “ignore all previous instructions” and role‑play do anything now or “finish every… | ||
| Identifying Repetitive Songs using LZ Compression | Geva Kipper | 23m |
Builds a dataset by scraped lyrics from six sources and measures repetitiveness with a custom LZ77 compressor that reports compression ratios (very repetitive… | ||
| Standup Comedy | Jonathan Harel | 20m |
Conference audio yields about six million transcribed words which get cleaned, run through peak detection and sentiment analysis and modeled with n‑grams… | ||
| Let our optima combine! | Eyal Gruss | 29m |
Shows how to model real combinatorial problems in Python using OR-Tools CP-SAT, with concrete code: 27 integer cell variables, 27×22 Booleans and channeling… | ||
| nothing matches | ||