We’ve built a template for fine-tuning Mistral 7B on Valohai. Mistral is an excellent combination of size and performance, and by fine-tuning it using a technique called LoRA, we can be very cost-efficient.

We’re thrilled to announce our new free trial for all aspiring ML pioneers! With the new free trial, we’ve made it easy to kickstart your journey with our handpicked templates.

ChatGPT continues to capture the public attention and many are looking to incorporate similar functionalities in their products. But is it a safe route for production-grade applications?

We've built a set of Hugging Face templates that make it super simple to use the latest and greatest in open-source ML. These templates are available through the Valohai Ecosystem.

With the popularization of LLM's developers and product folks are flocking to the space and testing out novel concepts. How will LLM products evolve over time?

The Valohai Ecosystem is a library of templates that enable users to kick off their projects with ease and reduce the amount of boilerplate code that needs to be written.

There's a new wave of automation being enabled by the combination of machine learning and smart devices. With the complexity of use cases and amount of devices increasing, we'll have to adopt MLOps practices designed for IoT and edge.

Product management is as massive a topic as machine learning so let's start with a fundamental question. When is it worthwhile to develop an AI product? A helpful tool most PMs have seen for this is the Sweet Spot for Innovation that IDEO popularized.

If you have a large team interested in large in-production use-cases, Metaflow is a great option for you. However, if you would like to standardize the whole ML workflow, MLflow is a viable option. In this article, you will learn about the similarities and significant differences between Metaflow and MLflow.

If you want a scalable platform that allows you to manage ML workflows, Metaflow is a better choice, while if you want to orchestrate a variety of tasks, Airflow is a viable option. In this article, you will learn about the key differences and similarities between Metaflow and Airflow.

Metaflow is a viable option if you need a specialized platform to build pipelines and SageMaker has a robust ecosystem through AWS. In this article, we will compare the major similarities and critical differences between Metaflow and Amazon SageMaker.

Finding the most suitable platform to build ML workflows may be a challenge. Some are looking toward specific tools built for ML/MLOps, such as Kubeflow, while others are looking at more general-purpose orchestrators such as Argo.

Creating a pipeline to automate ML workflows is necessary to save time and improve efficiency. There are two popular open-source tools for ML pipeline orchestration: Kubeflow and Metaflow. In this article, we will compare the differences and similarities between these two platforms.

Kubeflow and SageMaker have emerged as the two most popular end-to-end MLOps platforms. Kubeflow is the first entrant on the open-source side, and SageMaker has a robust ecosystem through AWS. In this article, we will compare the differences and similarities between these two platforms.

Databricks is a unified data analytics platform, while Kubeflow is an MLOps platform. In this article, we will look at how they are comparable and how they are very different.

Kubeflow and Airflow can both be used to orchestrate ML workflows. Airflow is the tool of choice for most engineers but this article will show what else MLOps platforms like Kubeflow can offer on top of DAGs.

Let me preface this article by saying there isn’t a single accepted definition of a machine learning lifecycle. Most articles about the machine learning lifecycle tend to focus only on a small portion of the actual lifecycle: the Experimentation loop.

DLOps, deep learning operations, is an evolution of MLOps, looking to answer the unique operational challenges that deep learning sets. A skeptic may look at it as unnecessarily muddying the waters with a new buzzword.

One of the unique aspects of Valohai is that despite being a proprietary platform it can run in fully private, even airgapped, environments. Why is this important? Machine learning often revolves around data that is sensitive and thus data security is a fundamental requirement.

The AIIA blueprint is an excellent starting resource for teams looking to implement their stack for machine learning development. The initiative draws inspiration from other popularized tech stacks.

As a data scientist or a machine learning engineer, you have probably heard about Kubeflow and MLflow. They are often compared against each other despite being quite different.

Putting together a suitable dataset for training a model can be one of the biggest challenges. Data augmentation is an approach where you start with an existing dataset and expand it to have more variety.

Summer is here and hopefully, for most of us, it means time to decompress. But if you are like me and learning is relaxing, podcasts are a great way to enjoy the summer weather while learning.

It's becoming more important to think about the competencies of a team rather than expecting every individual to be an expert at everything related to machine learning.

How can MLOps make consultant-client relationships more productive? Starting with machine learning is a massive, strategic undertaking, and many are turning to consultancies and contractors to take the first steps with AI.

March 16th, we held a webinar to follow up on our MLOps eBook. Together with our co-authors, we wanted to tackle the goal we set for MLOps in the eBook: “The goal of MLOps is to reduce technical friction to get the model from an idea into production in the shortest possible time to market with as little risk as possible.”

If you are involved with production machine learning in any way, understanding MLOps is essential. For people with software development experience, the easiest way to understand MLOps is to draw a parallel between it and DevOps.

The bus factor is a common term in software engineering describing the risk of a key contributor disappearing unexpectedly from a project – because they get hit by a bus. In machine learning the bus factor is magnified significantly.

Should a machine learning model be retrained each time new observations are available (or otherwise very frequently)? The answer is “it depends”, but this article looks at two components to consider: the use case and the costs.

Buying an MLOps platform is tricky and for that reason we’ve introduced a model where teams can sign up for a two-week proof-of-concept project to test out our platform with their environment and projects.

As you start incorporating machine learning models into your end-user applications, the question comes up: “When is the model good enough to deploy?” There simply is no single right answer.

To make it easier to consider what tools your organization could use to adopt MLOps, we’ve made a simple template that breaks down a machine learning workflow into components.

Machine learning and artificial intelligence allow businesses to gain new insights and improve their business processes. However, they expose companies to additional risks because humans do not explicitly program the algorithms. Let's look at some of these risks and how data scientists and compliance officers can help mitigate them.

Using the MLOps platform allows you to manage everything about machine learning in production, where each new update doesn’t feel like an entirely new project and easily dovetails to the last.

To make an analogy to a more traditional industry, machine learning is shipping goods while MLOps is containerization. And much like containerization of global shipping, MLOps is equal parts process and infrastructure.

For a long time, most machine learning initiatives have been stuck in a persistent state of proofs-of-concept. However, in the past year, we’ve seen a rapid acceleration of machine learning models getting real-world use. Consequently, machine learning engineers are increasingly sought after – nearly catching up to data scientists in posted jobs.