A developer shares talks on building safe AI agents for high-stakes industries using Go and durable execution, and announces an upcoming meetup.
A tutorial on creating a production-ready Docker image for PyTorch models using Torch Serve, including model archiving and dependency management.
A guide on transitioning Generative AI applications from proof-of-concept to production, covering architecture, security, and operations.
A reflection on how LLMs have simplified AI prototyping compared to traditional ML, but may lead to similar deployment disappointments.
Interview with Itai Bar Sinai, co-founder of Mona Labs, discussing AI monitoring, product-oriented data science, and the Israeli ML community.
Analyzes common pitfalls in AI adoption, arguing that technical and product maturity models can hinder practical implementation.
An update on how Monzo integrated machine learning across its organization in 2022, covering team structure, growth, and new initiatives.
Final part of a series proposing a research agenda for ML monitoring, focusing on data management challenges like metric computation and real-time SLI tracking.
Explores the challenges of using Prometheus for ML pipeline monitoring, highlighting terminology issues and technical inadequacies.
A workshop series on using Hugging Face Transformers with Amazon SageMaker for enterprise-scale NLP, covering training, deployment, and MLOps.
Analyzes post-deployment ML issues and categorizes them to advocate for better monitoring tools, using Zillow's case as an example.
A podcast interview discussing common ML mistakes, quantifying impact, and career growth for machine learning engineers.
A critique of current ML monitoring tools and a proposal for rethinking evaluation in streaming machine learning systems.
Learn how to integrate the Hugging Face Hub as a model registry with Amazon SageMaker for MLOps, including training and deployment.
Guide to building an end-to-end MLOps pipeline for Hugging Face Transformers using Amazon SageMaker Pipelines, from training to deployment.
A talk on system design principles for building production recommendation systems and search engines, presented at an MLOps Community meetup.
Introducing mltrace, an open-source lineage and tracing tool for debugging and maintaining production machine learning pipelines.
Explores the concept of feature stores in machine learning, presenting a hierarchy of needs from basic access to full automation.
The article argues that the choice of machine learning library (like PyTorch or TensorFlow) is less critical than building robust data and production pipelines.
Introducing efsync, an open-source MLOps toolkit for syncing dependencies and model files to AWS EFS for serverless machine learning.
