A chain is as strong as its weakest link - and your AI/ML infrastructure is only as fast as your slowest component. If you train machine learning models with GPUs, then your weak link may be your storage solution. The result is what I call the “Starving GPU Problem.”  The Starving GPU problem occurs when your network or your

Explore the essential role of Data Engineers in unleashing the true power of AI! Data Engineers have a critical foundation in cleaning and structuring raw data for ML success. Learn why their expertise in data infrastructure, feature engineering, and pipeline optimization is indispensable.

Much has been said lately about the wonders of Large Language Models (LLMs). Most of these accolades are deserved. Ask ChatGPT to describe the General Theory of Relativity and you will get a very good (and accurate) answer. However, at the end of the day ChatGPT is still a computer program (as are all other LLMs) that is blindly executing

Over the past few months, I have written about a number of different technologies (Ray Data, Ray Train, and MLflow). I thought it would make sense to pull them all together and deliver an easy-to-understand recipe for distributed data preprocessing and distributed training using a production-ready MLOPs tool for tracking and model serving. This post integrates the code I presented

Most machine learning projects start off as a single-threaded proof of concept where each task is completed before the next task can begin. The single-threaded ML pipeline depicted below is an example. However, at some point, you will outgrow the pipeline shown above. This may be caused by datasets that no longer fit into the memory of a single process.

The calendar year 2023 will be a meaningful one, perhaps one of the most meaningful ones, when the history of AI is written. It was, in essence, the big bang.  It started in late 2022 with OpenAI’s ChatGPT but it was the response that was so breathtaking. Within months we had Meta’s LLaMA 2, Google’s Bard chatbot

Introduction Distributed data processing is a key component of an efficient end-to-end distributed machine-learning training pipeline. This is true if you are building a basic neural network for statistical predictions where distributed training could mean each experiment runs in 10 minutes vs. an hour. It is also true if you are training or fine-tuning a Large Language Model (LLM) where

This post was written in collaboration with Amit Kesarwani from lakeFS. The reality of running multiple machine learning experiments is that managing them can become unpredictable and complicated - especially in a team environment. What often happens is that during the research process, teams constantly change configuration and data between experiments. For example, try several training sets and several hyperparameter

Introduction Generative AI represents the latest technique an enterprise can employ to unlock the data trapped within its boundaries. The easiest way to conceptualize what is possible with Generative AI is to imagine a customized Large Language Model - similar to the one powering ChatGPT - running inside your firewall. Now, this custom LLM is not the same as the

A lot of ink has been spilled on the significance of the AI/ML technology wave (here are our posts). What doesn’t get attention, but probably should, is how AI/ML is remaking the technology power structure inside the enterprise. As companies reorganize around a data-centric orientation, they are also reorganizing who makes and executes the technology architecture. While

Introduction In a previous post, I presented feature extraction, which is a technique for utilizing pre-trained Large Language Models (LLMs) to solve a custom problem without having to retrain the model. Feature extraction is one of two ways to use the knowledge a model already has for a task that is different from what the model was originally trained to