# Data quality

128 sessions · follows the tags data-quality, data-pipelines
Page: https://mlopstalks.com/threads/data-quality

The oldest complaint in the archive, restated for every new kind of model.

## 2020

19 sessions.

- [Mid-Scale Production Feature Engineering](https://mlopstalks.com/talks/mid-scale-production-feature-engineering) (Dr. Venkata Pingali, Scribble Data). Claim: Dr. Venkata Pingali says model reproducibility is insufficient without reproducibility and lineage for the data used by the model. [13:13](https://www.youtube.com/watch?v=1CcYuVVwOGg&t=793s)
- [TrueLayer's MLOps Pipeline](https://mlopstalks.com/talks/truelayers-mlops-pipeline) (Alex Spanos, TrueLayer). Pushed back: Alex Spanos says the machine learning pipeline should eventually resemble mature DevOps practice, while noting that machine learning has additional moving parts such as data versioning, parameters, and metrics. [51:03](https://www.youtube.com/watch?v=cUxK28ocZcw&t=3063s)
- [10 Years Deploying ML in the Enterprise: The Inside Scoop!](https://mlopstalks.com/talks/10-years-deploying-ml-in-the-enterprise-the-inside-scoop) (Charles Martin, MLOps Community). Claim: Charles Martin says that unvalidated and changing data inputs make it difficult to automate machine learning systems reliably. [29:06](https://www.youtube.com/watch?v=WmZL1AQPbOw&t=1746s)
- [Monitoring the Machine Learning Stack](https://mlopstalks.com/talks/monitoring-the-machine-learning-stack) (Lina Weichbrodt, DKB). Pushed back: Lina Weichbrodt said that real-time response monitoring is needed in addition to offline data-quality checks such as Great Expectations or TensorFlow Data Validation. [21:00](https://www.youtube.com/watch?v=Un30yb1WlpU&t=1260s)

15 more from 2020 on this thread: https://mlopstalks.com/threads/data-quality/2020

## 2021

23 sessions.

- [The revolution of Federated Learning](https://mlopstalks.com/talks/the-revolution-of-federated-learning) (Fabiana Clemente, MLOps Community & Ramen Dutta, TensoAI). Claim: Successful adoption of federated learning requires high-quality data, talented data scientists, and the ability to adapt to new data sets and quickly validate and release models. [15:46](https://www.youtube.com/watch?v=qRBfftNLiDQ&t=946s)
- [Machine Learning Feature Store Panel Discussion](https://mlopstalks.com/talks/machine-learning-feature-store-panel-discussion) (Daniel Galinkin, iFood & Matias Dominguez, Rappi & Simarpal Khaira, Intuit). Claim: Matias Dominguez says a small company without a market-validated product may not need to buy or build a full feature store. [9:29](https://www.youtube.com/watch?v=-TGp2qKz8tA&t=569s)
- [ML Tests](https://mlopstalks.com/talks/ml-tests) (Svet Penkov, Efemarai). Pushed back: Svet Penkov takes the view that measuring data quality beyond basic validity is not always meaningful and that model performance in the intended domain is the more useful measure. [29:34](https://www.youtube.com/watch?v=c59Wmg2-jpA&t=1774s)
- [Building 12-Factor Data Apps with Kedro](https://mlopstalks.com/talks/building-12-factor-data-apps-with-kedro) (Ivan Danov, QuantumBlack). Pushed back: Ivan Danov says Kedro is not another orchestrator like Airflow or Kubeflow because it focuses on pipeline authoring rather than workflow execution and monitoring. [39:06](https://www.youtube.com/watch?v=j-qPZkdJREw&t=2346s)

19 more from 2021 on this thread: https://mlopstalks.com/threads/data-quality/2021

## 2022

17 sessions.

- [Trustworthy Data for Machine Learning](https://mlopstalks.com/talks/trustworthy-data-for-machine-learning) (Chad Sanderson, Convoy). Pushed back: Chad Sanderson initially argued that data scientists should own the quality of the modeling code they write, but later concluded that data scientists are not software engineers and that engineers should own data quality. [42:26](https://www.youtube.com/watch?v=k98bPYlZXds&t=2546s)
- [DataOps is a Software Engineering Challenge](https://mlopstalks.com/talks/dataops-is-a-software-engineering-challenge) (Micha Kunze, Maersk). Pushed back: Micha Kunze argues that full-job or pipeline tests are often more valuable and stable than testing every individual function, although he still uses unit tests for some complicated transformations. [51:23](https://www.youtube.com/watch?v=iAr6woT1oIQ&t=3083s)
- [The Post Modern Stack](https://mlopstalks.com/talks/the-post-modern-stack) (Jacopo Tagliabue, Coveo). Pushed back: Jacopo Tagliabue disputes the view that building an ML pipeline requires a very large team or a million people. [9:20](https://www.youtube.com/watch?v=JUFgKSZo1j4&t=560s)
- [Labeled Datasets that Correct Themselves Automatically](https://mlopstalks.com/talks/labeled-datasets-that-correct-themselves-automatically) (Curtis Northcutt, Cleanlab). Pushed back: Cleanlab's results are not based only on model mistakes, because the model also reflects errors in the data it was trained on. [29:55](https://www.youtube.com/watch?v=IwDGDAHgzAY&t=1795s)

13 more from 2022 on this thread: https://mlopstalks.com/threads/data-quality/2022

## 2023

21 sessions.

- [ML in Production: A DS from Ubisoft Perspective](https://mlopstalks.com/talks/ml-in-production-a-ds-from-ubisoft-perspective) (Jean-Michel Daignan, Ubisoft). Claim: Jean-Michel Daignan says scalability testing can be more important for his pipelines than unit-testing every individual function. [16:06](https://www.youtube.com/watch?v=6mea_qHJLkw&t=966s)
- [Multilingual Programming and a Project Structure to Enable It](https://mlopstalks.com/talks/multilingual-programming-and-a-project-structure-to-enable-it) (Rodolfo Núñez, Entel). Claim: Development pipelines can call scripts in sequence or in parallel, record timings and logs, and give production engineers a simple command to run the required workflow. [43:03](https://www.youtube.com/watch?v=sAf7ycA1k8s&t=2583s)
- [Dataframes Are All You Need: MLOps on Easy Mode](https://mlopstalks.com/talks/dataframes-are-all-you-need-mlops-on-easy-mode) (Jay Chia, Eventual). Claim: Jay Chia says dataframes can support much of the MLOps workflow, including reading and writing data, exploring and processing it, feeding training pipelines, evaluating models, and running batch predictions. [15:07](https://www.youtube.com/watch?v=o5SVCJcjrM4&t=907s)
- [Building LLM Applications for Production](https://mlopstalks.com/talks/building-llm-applications-for-production) (Chip Huyen, Claypot AI). Claim: Chip Huyen says LLMs do not reliably follow a required output schema, making it difficult for applications to parse their responses. [5:19](https://www.youtube.com/watch?v=spamOhG7BOA&t=319s)

17 more from 2023 on this thread: https://mlopstalks.com/threads/data-quality/2023

## 2024

37 sessions.

- [The Rise of Modern Data Management](https://mlopstalks.com/talks/the-rise-of-modern-data-management) (Chad Sanderson, Gable). Pushed back: Chad Sanderson says the main problem with data lineage is upstream of analytical systems, rather than downstream inside tools such as Snowflake or Databricks. [29:07](https://www.youtube.com/watch?v=h5agcL6t4uM&t=1747s)
- [ML and AI as Distinct Control Systems in Heavy Industrial Settings](https://mlopstalks.com/talks/ml-and-ai-as-distinct-control-systems-in-heavy-industrial-settings) (Richard Howes, Metaformed). Pushed back: Richard Howes rejects the idea that subject-matter professionals should be replaced in large-scale ML analysis and says they are needed to validate the outputs. [38:35](https://www.youtube.com/watch?v=wQJBuA3Hm30&t=2315s)
- [How Feature Stores Work](https://mlopstalks.com/talks/how-feature-stores-work) (Simba Khadder, Featureform). Pushed back: Simba Khadder argues that data scientists should not have to become expert data engineers to build production-grade feature pipelines. [8:27](https://www.youtube.com/watch?v=DQWjPEXdzJM&t=507s)
- [Building an ML Platform from scratch](https://mlopstalks.com/talks/building-an-ml-platform-from-scratch) (). Pushed back: Ben says feature transformations should ideally be centralized in SQL mesh or a similar modeling system, while Eric favors monolithic pipelines early and sees decoupling as a later maturity step. [1:26:27](https://www.youtube.com/watch?v=9LxsViICsJo&t=5187s)

33 more from 2024 on this thread: https://mlopstalks.com/threads/data-quality/2024

## 2025

9 sessions.

- [Look At Your ****ing Data 👀](https://mlopstalks.com/talks/look-at-your-ing-data) (Kenny Daniel, Hyperparam). Pushed back: Kenny argued that the lack of discussion about LLM data quality is not explained only by data sources being secret; researchers also rarely discuss how they clean and filter data. [4:13](https://www.youtube.com/watch?v=6EMnkAHmoag&t=253s)
- [Streaming Ecosystem Complexities and Cost Management](https://mlopstalks.com/talks/streaming-ecosystem-complexities-and-cost-management) (Rohit Agrawal, Tecton). Claim: A typical streaming pipeline connects Kafka to a processor such as Spark or Flink, then to storage and a serving layer, with each part requiring different skills. [5:11](https://www.youtube.com/watch?v=0axFbQwHEh8&t=311s)
- [How Sama is Improving ML Models to Make AVs Safer](https://mlopstalks.com/talks/how-sama-is-improving-ml-models-to-make-avs-safer) (Duncan Curtis, Sama). Claim: Human work is usually the most expensive part of an AI data pipeline, except in specialized cases where data collection involves expensive equipment. [4:35](https://www.youtube.com/watch?v=UJ6uA46yWzM&t=275s)
- [GraphBI: Expanding Analytics to All Data Through the Combination of GenAI, Graph, & Visual Analytics](https://mlopstalks.com/talks/graphbi-expanding-analytics-to-all-data-through-the-combination-of-genai-graph) (Paco Nathan, Senzing). Pushed back: Weidong Yang disputes treating ontology as universally valid and says its truth must be limited to a defined domain. [21:00](https://www.youtube.com/watch?v=2i8hW4O1WVs&t=1260s)

5 more from 2025 on this thread: https://mlopstalks.com/threads/data-quality/2025

## 2026

2 sessions.

- [Fixing GPU Starvation in Large-Scale Distributed Training](https://mlopstalks.com/talks/fixing-gpu-starvation-in-large-scale-distributed-training) (Kashish Mittal, Uber). Pushed back: Kashish Mittal rejects the idea that low GPU utilization is mainly caused by an expensive model forward pass and says the data pipeline is the problem. [9:08](https://www.youtube.com/watch?v=1WFffCGhm7U&t=548s)
- [Why AI Agents Shouldn't Replace Your Fraud Models](https://mlopstalks.com/talks/why-ai-agents-shouldnt-replace-your-fraud-models) (Varant Zanoyan, Zipline AI). Claim: In agentic experimentation, an agent can create features, add them to a model, train it, evaluate the results, and deploy an end-to-end pipeline to development. [7:38](https://www.youtube.com/watch?v=HaWk8kAD8ZU&t=458s)