hayaData 2024
as the organisers put it
hayaData conference is a not-for-profit event built for the data community, data engineers & scientists, BI & data analysts, developers, researchers, and everyone interested in big data!
| title | author | length |
|---|---|---|
| hayaData 2024 Highlights | 01m | |
A fast-paced reel that stitches keynote excerpts and session clips from hayaData 2024, focusing on concrete demos and outcomes rather than high-level hype… | ||
| Opening remarks | Orit Mantzu | 16m |
Community conference reports rapid growth: 1000+ attendees, 26 partners, 7 communities and expansion from two to 4 tracks delivering 45 talks selected from 300… | ||
| The Future of Data is Words | Josef Goldstein | 30m |
SQL celebrates 50 years while LLMs push natural-language queries; demos show Flink SQL stream pipelines that convert multi-currency order streams into unified… | ||
| Investing in Data: Unlocking Tomorrow's Opportunities | Maayan Manela, Boaz Morris, Helen Ani Renana Ashkenazi | 28m |
Four investors argue that winning data companies turn raw collection into cash by delivering meaningful insights not piles of files, and by proving early… | ||
| Sparkless Patterns: The New Data Stack and Modern Data Architecture | Alon Agmon | 30m |
Claims that DuckDB Polars and Apache DataFusion let small nodes replace heavy Spark clusters with concrete results: a Lambda running DuckDB reads Iceberg on S3… | ||
| Boosting Cost Efficiency by Cost Aware Architecture | Doron Hoffman | 17m |
Claims a 40% cost reduction by changing storage layout and instance choice, with concrete numbers: analysis of cost per TB-month and cost per million writes… | ||
| Personalizing User Content - The Power of a Vector & Search DB | Shelly Glick | 26m |
Taboola replaces offline candidate lists with a Vespa-based vector search that ranks millions of items into a few thousand candidates in ~300ms, using Debezium… | ||
| Ooops... I Deleted the Whole Production Table. What Did I Learn From It? | Yerachmiel Feltzman | 06m |
At 6pm a mistaken rm removes an entire production table with no backup; the team builds a modular backfill pipeline that bypasses CDC, scans the source and… | ||
| From Scaling APIs to Scaling Data to Scaling Your Team | Shani Aharoni | 07m |
Starts with a 7am failed ETL job that exposes missing history and costly re-runs, and argues that hiring backend engineers accelerates building reliable… | ||
| Congratulations! Meet Your New Little Brother | Ronit Nehemia | 06m |
Ministry of Education launches a 10-unit major in data analysis now taught in about 100 high schools across Jewish, Arab, Bedouin, Druze and religious sectors… | ||
| It’s Just an Alert, Don’t Wake Up! | ALON NATIV | 06m |
Defines a concrete KPI: percentage of alerts that require immediate, actionable work, with a practical goal of zero false positives by design. Uses percentiles… | ||
| Building a Data-Driven Culture: An Anthology of Strategies and Insights | Mor Nitzan | 14m |
Built a production lead-scoring ML model for a B2B sales org, runs A/B tests and shows higher conversion but initially fails because reps’ bonuses reward speed… | ||
| Live in the Data Wild West: The Data Contracts Sheriff | Tal Peretz | 27m |
Claims a concrete platform that enforces data contracts across ingestion, streaming and analytics, handling schema evolution and preventing bad writes into… | ||
| Dropping Doesn’t Mean Losing | Daniel Gur | 15m |
Outbrain runs on Kubernetes and handles about 5 million requests/sec, 200M metrics/min and ~15 TB of logs/day, so a lightweight rate-control layer opens… | ||
| Exploring the Depths of Apache Iceberg's Metadata Capabilities | Amit Gilad | 13m |
Explains how Iceberg’s metadata tables actually solve performance and operational problems: use the snapshots as git commits table to get per-write metrics… | ||
| Unifying Real-Time and Data Lake: Yotpo's Transition from Chaos to Coherence | Victor Perepelitsky | 29m |
Yotpo replaces costly daily DB snapshots with CDC with Debezium feeding Kafka, builds tables in the lake via streaming, and enforces contracts (Avro + rules)… | ||
| Big Data Processing with GPUs | Ilay Chen | 14m |
PayPal reports 70% cost reduction by moving a Spark 2 pipeline (140 nodes, 2.0 h) to Spark 3 with GPU acceleration (30 nodes, 1.3 h) using Spark RAPIDS and two… | ||
| The Hitchhiker's Guide to Advanced A/B Testing Techniques | Allon Korem | 28m |
Covers concrete checks and fixes: run a sample ratio mismatch test and stop if allocation deviates (planned 50k users, saw ~20k so far, initial split 10k vs… | ||
| Democratizing Data Pipelines: Introducing the AI-Powered ETL Assistant | Haya Axelrod Stern | 16m |
A working ETL assistant that uses hundreds of existing pipeline examples plus a knowledge upload to an LLM to translate natural-language requests into cron… | ||
| Metric Store | Ben Hababo, Mickey Rozen | 27m |
Company standardizes analytics into a four-layer pipeline (L1 ingest → L2 validated → L3 curated → L4 metrics) and builds a metric store as wide denormalized… | ||
| Data Driven Advocacy: Using Vizualization Skills for Hasbara | Nir Smilga, Omer Biber | 16m |
Used Tableau plus custom PDF extraction to build interactive timelines and a dashboard comparing UN condemnations to civilian deaths, drawing on 21 countries… | ||
| Beyond Empowerment: Data Accuracy Challenges in Self-Service BI | Maria Yanson | 12m |
A lightweight validation system marks trusted dashboards with a BI Validated logo by matching dashboard URLs against a Google Sheets whitelist, synced into… | ||
| Battle of the Titans: Python vs. Tableau in Live EDA | Shuki Cohen, Efrat Garber Aran | 31m |
In the rapidly evolving field of data analytics, the choice between mastering Python or Tableau remains a prevalent dilemma among professionals. This session… | ||
| Here, There and Everywhere: Spotting Data Opportunities 101 | Mor Hananovitz | 14m |
Factory line metric rises from 70% → 95% in 2 months after a $10 departmental UI that prioritizes parts for workers; solution uses simple CSV automation… | ||
| Unlocking the Power of AI for Seamless Self-Service BI Transfer | Eyal El Bahar | 17m |
Lightricks shifts to AI-first product and BI, building three projects: a pre-commit parser called comic that auto-generates docs and saves ~2–3 minutes per PR… | ||
| Threat Hunting Powered by Efficient and Straightforward Anomaly Detection on Your Data Lake | Ori Nakar | 24m |
Serverless SQL pipeline reads each source table once, aggregates JSON-defined counters into dedicated tables and flags anomalies with simple rules (value >… | ||
| Dating with a Super Model: Why Good Prompt Engineering for Data Monitoring Requires Some Flirting | Reut Vilek | 13m |
Builds an automatic data-monitoring pipeline for tens of use cases and reports three concrete lessons: large generative models often hallucinate or bluntly… | ||
| Evaluating the Unseen: Supervised Evaluation for Unsupervised Algorithm | Ben Harel | 26m |
Uses a supervised edge classifier to label account pairs so clustering quality can be measured with v-measure instead of relying on silhouette scores. Features… | ||
| Tailor-Made LLM Evaluations: How to Create Custom Evaluations for Your LLM | Linoy Cohen | 14m |
Compares two practical evaluation paths: benchmarking using task-specific datasets to catch training regressions and contamination, and LLM-as-a-judge using… | ||
| Correlation Sucks - Long Live Causality! Mastering Causal Inference in Data Science | Amit Sasson | 28m |
Years of R&D or a seasonal ad campaign require causal, not just predictive, answers: claim is that A/B tests give unbiased effects when feasible, otherwise… | ||
| Learning the Ropes of Synthetic Data | Noa Zamstein, PhD | 14m |
A practical primer on synthetic data that shows how models estimate a probability density, resample to produce avatars, and preserve population-level utility… | ||
| LLMs and Knowledge Graphs: A Case Study in Blue(y) | Stav Shamir | 14m |
A working Bluey proof‑of‑concept builds a Neo4j knowledge graph of characters, episodes and extracted concepts, uses an LLM to extract three concept types per… | ||
| Beyond Opt-Outs: Quantifying User Notification Harm | Orna Amir, PhD, Hila Kantor | 25m |
Presents a measurable framework that predicts notification harm by combining feature engineering, ML models (Lasso and Random Forest), free‑text clustering… | ||
| Navigating the Uncharted: Ensuring Prompt Quality in the Age of Language Models | Ortal Ashkenazi | 16m |
A production prompt that summarized calls begins hallucinating on short calls and surfaces false actions, so an iterative process uses manual QA (content… | ||
| Redefining Agile for Data Science: A Team First Approach | Andres Asaravicius | 16m |
Argues that Agile works for data science if retooled around team practices rather than individual silos, showing a practical framework used across several… | ||
| I Want to Build a RAG System, Now What? | Roy Miara | 30m |
Rapidly prototype a RAG POC: chunk documents into 400-token chunks, embed with an encoder, store in a vector DB (Pinecone/Canopy noted by uploader), retrieve… | ||
| Transforming Medical Records Into a Heart Attack Prediction Model | Shiri Gaber | 17m |
Hello Heart combines daily app data (blood pressure, meds, symptoms) with EHRs, builds a daily ETL that extracts and aggregates hundreds of thousands of raw… | ||
| Discover the Modern Way to Build Real-Time Analytics Solutions at Any Scale | Golan Nahum, Ilya Gulman | 32m |
Shows a concrete before/after: a startup pipeline built from Postgres/Mongo/Elasticsearch + Kafka + Spark + S3 + Iceberg ends with data islands, heavy sync… | ||
| Accelerate Data Pipeline Development Time: Learn How to Unleash Rivery’s Gen AI for a 20X Leap | Aviv Noy | 14m |
Rivery’s Gen AI uses retrieval-augmented generation to read API docs, infer auth and parameters, and auto-generated connector code plus schema and config that… | ||
| Building Data AI Application with the Snowflake AI Data Cloud | Roy Pasternak | 30m |
Company use case builds a customer-facing data+AI app on Snowflake for a fintech, keeping one copy of data and avoiding copies for Spark teams. Architecture… | ||
| A Brave New World: How New Lineage Technology Can Reduce Data Spend 40% | Guy Biecher | 14m |
Claims a multi-layered lineage system that overlays dbt models, Snowflake tables and Tableau dashboards with usage and cost metrics to map which producers feed… | ||
| The 10 Commandments for Crafting Custom LLMs with Domain Mastery | Udi Menkes | 28m |
Focuses on turning proprietary data and domain expertise into deployed LLMs using an internal genos platform and a cross‑product assistant called in assist… | ||
| How Similarweb Serves 100s of TBs to Their Worldwide Users in Milliseconds | Yoav Shmaria | 17m |
Similarweb serves hundreds of TBs of web data with consistent millisecond latency by combining key‑value fast paths for simple time‑series with a columnar… | ||
| How to Decipher User Uncertainty with GenAI and Vector Search | Ben Greenberg | 30m |
Shows how BERT and GPT embeddings plus vector indexes turn vague queries into semantic matches, not keyword hits, and compares distance metrics (Euclidean… | ||
| nothing matches | ||