Blog

As AI workloads grow in complexity and scale, the challenge has become getting them into production, reliably and efficiently. The traditional stack was built for ETL, not AI.

Distributed computing is one of those “good problems to have.” It means your ML models, AI pipelines, or simulations are big enough that a single machine just won’t cut it! But the question quickly becomes: How do you actually scale an AI workflow?

Flyte 2.0 represents a fundamental shift in workflow orchestration, moving from static DAG composition to truly dynamic workflow execution.

Nobody likes paying taxes. But what’s the worst kind? The one you don’t even know you’re paying.

Getting an AI application to behave the way you want takes more than just wiring things together. The real effort is in the setup: testing prompts, juggling models, tweaking parameters, often for days until you land on something that works reliably.

Over the past decade, AI infrastructure has evolved dramatically. We’ve moved from static ETL pipelines to dynamic machine learning workflows, and now to intelligent, autonomous agents. 

SLMs help bridge the experiment-to-production gap, but only if you use them the right way.

Leaders from Artera AI and Union.ai came together to discuss how top healthcare and biotech teams build scalable, compliant AI infrastructure. Learn best practices in workflow orchestration in healthcare, data locality, security, and global AI deployment.

Amazon S3 Vectors is the first cloud object store with native support for storing and querying vectors. That’s a big deal; no need for a separate vector database if you’re already on AWS.

Modern AI workloads are growing in complexity, scale, and cost. As AI and ML teams build longer-running workflows and agents, AI orchestration becomes an essential tool in the implementation of their systems.

This article charts the most pressing problems that orchestration can solve, and introduces the vision for Flyte 2.0. We’ll publish technical deepdives on Flyte’s capabilities over coming weeks.

Let’s chart the history of orchestration and explore why it has become a non-negotiable for the future of AI and agents.

At Union.ai, our philosophy is simple: provide you with every tool needed to unlock maximum value from your AI infrastructure.

Bringing transparency and control to every stage of your model and app lifecycle.

As geospatial machine learning engineers, we often face a common challenge.

When building real-world AI systems, performance isn’t just about how fast your models run when deployed — it's also about how efficiently your infrastructure supports them.

SEATTLE, WA— Union.ai, the developer suite for orchestrating, training, and serving AI models, today announced that it has achieved the Google Cloud Ready - BigQuery Designation. 

In this fireside chat, we talk with Arize’s Head of Developer Relations John Gilhuly about LLM Observability: Tracing, Evaluations, and Real-Time Insights.

In AI and machine learning, the need for reproducibility is essential for ensuring reliability, transparency, and trustworthiness of models and experiments.

Imagine if building compound AI apps was as easy as piecing together Legos. At Union, we're abstracting the infrastructure layer to make this a reality.

As machine learning systems become deeply embedded in software infrastructure, their security vulnerabilities pose critical risks.

Read on for new features, upcoming events, and job openings. 

In the fast-paced world of machine learning, data scientists and researchers are constantly battling infrastructure challenges that slow down innovation and drain organizational resources.

Protein folding has profound implications for understanding disease, drug development, and fundamental life processes.

Read on for a re:Invent recap, new features, upcoming events, and job openings.

Building scalable machine learning projects is challenging enough without the added burden of hidden or unpredictable infrastructure costs.

AI workflows often use containers as isolated environments per task, but each spin-up requires initializing dependencies, configuring models, and loading data, slowing down workflows that need rapid, repeated executions.

Developing and deploying AI models at scale is challenging. Many teams face obstacles like disconnected workflows, runaway or prohibitive infrastructure costs, and slow time-to-market for AI solutions.

Visit Union in the re:Invent Expo Hall as part of the NVIDIA booth #1620.

Reaching the amazing milestone of 50 million downloads, we asked Niels Bantilan, Pandera’s creator, to reflect on the journey of getting here and what might be next.

Reproducibility is the cornerstone of science and engineering: Without it, we can’t reliably build on top of each other’s findings and work.

So, I had this idea the other day: What if I could get ChatGPT-like AI to run on my iPhone and speak to me in my native language, Telugu? Not through some API or cloud service, but actually running directly on my phone. 

Ollama is a platform designed to simplify running open-source large language models (LLMs) locally.

Flyte is back again for another exciting round at this year’s Hacktoberfest!

I’m excited to announce that David Jakubowski has joined Union as President.

In this fireside chat, we talk with Groundlight co-founder and CTO Leo Dirac about the future of computer vision and robotics. 

At Union, we understand the complexities of deploying generative AI in production.

In the machine learning (ML) and artificial intelligence (AI) domain, managing, tracking, and visualizing model training processes, especially at scale, is a significant challenge. 

With Flyte’s latest plugin for Weights & Biases, you can now effectively run Machine Learning or AI workflows on Union and integrate with Weights & Biases capabilities.

It’s been a journey. There have been users successfully running Flyte on Azure but, unfortunately, nothing publicly documented.

We are excited to announce that Union Serverless users can now harness the power of NVIDIA A100 Tensor Core GPUs and NVIDIA L4 Tensor Core GPUs.

NVIDIA Parabricks is a software suite that accelerates genomic sequence analysis by reimplementing industry standard tools to use the parallel...

Pandera v0.20.* now supports Pyarrow data types in the pandas validation engine 🚀.

Bioinformaticians and ML engineers often need to balance performance versus cost when running data-intensive workloads.

Orchestrate your AI, with serverless execution and pay-as-you-go pricing.

OpenAI real-time client is ideal for real-time responses, especially for use cases and needs where price is not a factor...

Artifacts provide a core abstraction that serves as an interface between the different teams that work together to build AI-powered products.

Batch workloads often involve large data processing and dependency on other batch workloads.

Machine learning engineering is a complex process, spanning data procurement and processing, model training, deployment, and scaling.

We have had the privilege of seeing data teams experience the value of a unified platform for both machine learning and data pipelines.

The day is finally here! Pandera 0.19.0 ships support for Polars.

Zilliz provides enterprise-grade AI technologies, including one of the world’s most popular open-source vector databases, Milvus.

AI video dubbing or translation has surged in popularity, breaking down language barriers and enabling communication across diverse cultures.

As generative AI models become more capable and deployed in various contexts, optimizing the model in terms of throughput and memory consumption...

Today’s AI development lifecycle is marked by multiple teams collaborating to create AI-driven products.

Here at Union, we’ve been hard at work enhancing the FlyteConsole.

After an electrifying NVIDIA GTC 2024, I’ve been back in Seattle going through my notes, photos, and a whirlwind of memories!

Explore an end-to-end workflow: fine-tuning SAM, batch prediction, user approval, and deployment on SageMaker

In modern data-driven and machine-learning workflows, efficient orchestration & execution of tasks are crucial in achieving productivity & scalability

Collaboration underscores Union's commitment to democratize AI application development

Traditionally, organizations have selected a single computing provider on which to base their operations.

Pandera 0.18 introduces two new configuration settings that control how validation happens.

Flyte is a workflow orchestrator that unifies machine learning, data engineering, & data analytics stacks for building robust & reliable applications.

This is part 3 of a blog series on the need for end-to-end reproducibility in AI development.

NVIDIA's GTC 2024 is just around the corner, and tech enthusiasts, industry leaders, and innovators worldwide are gearing up!

In a previous post we highlighted the need for reproducibility in AI development, and explored some of the challenges of achieving this...

We’re excited to announce that the Union platform is now available on the Google Cloud Platform (GCP) Marketplace.

The recent explosion of interest in AI has sparked a surge of investment aimed at bringing the technology into virtually every consumer and...

Developing AI workflows locally, then deploying to remote environments can lead to problems and slow downs.

Hello, Flyte community! The Union team behind Flyte is thrilled to announce the Flyte Community Raffle! 

Voxel51 is the data-centric machine learning software company behind FiftyOne, an open-source toolkit for building computer vision workflows.

Flyte is an AI orchestration platform for rapid Data, ML, & AI development, particularly useful for scaling pipelines in production environments.

Pryce recently joined the Union team as a Solutions Architect.

We’re excited to open the Union platform to AWS developers as it is now available in the AWS Marketplace.

By walking through a genomic alignments code and a Streamlit app, explore how Union makes it easier to connect external inputs to pipelines.

What’s It Like Working at Union.ai?

It’s KubeCon season, and cloud-native enthusiasts will flock to Chicago November 6-9 to hear the latest innovations to all things Kubernetes. 

I’m super excited to announce that Pandera now supports pydantic v2! 🎉

Union.ai is thrilled to attend and sponsor the MLOps World Conference for the second year in a row!

When NGS applied the power of massively parallel processing to the analysis of DNA, it transformed the biotech playing field.

Graphics Processing Units have roots deep in gaming, but they’re serious business when it comes to machine learning.

LLMs for data cleaning: Yay or nay?

Kubernetes guru Kelsey Hightower spoke with Union.ai about the business case for LLMs and AI and their potential to reshape the workplace. 

Do more with less: Refine the 70 billion parameter Llama 2 model on your dataset with a bunch of T4s

Large Language Models (LLMs) are this season’s hot technology topic, but what will it take to establish their lasting success in the enterprise?

A 5-minute cheatsheet outlining the challenges and solutions for building Large Language Models for production.

Friday afternoon, you’re browsing the website of your usual event-ticketing portal trying to find something interesting to do this weekend.

Artificial intelligence pipelines are infrastructure intensive. They may entail machine- or deep-learning models that are more iterative...

I’m super excited to announce the availability of Pandera 0.16! This release features a suite of improvements and bug fixes.

Large language models (LLMs) have taken the world by storm, revolutionizing our understanding and generation of human-like text.

Union.ai is excited to share that we will be attending and sponsoring the Toronto Machine Learning Summit conference happening from June 12-13.

Imagine you have a dataset of thousands of audio files that need to be transcribed.

The founder of VC Nava Ventures describes the big vision he sees for Union.ai.

Union Cloud enables any organization to operationalize complex AI with data ownership, governance and cost efficiency.

When to manipulate the input prompt and when to roll up your sleeves and update parameter weights.

The Union.ai team kept busy last week at PyData Seattle 2023.

We’re excited to announce that Union Cloud has achieved the SOC2 Type II Certification for Security, Availability and Integrity.

Every year, PyData Seattle brings together experts, practitioners, and enthusiasts in the field of data science and machine learning.

The World’s Most Important Gathering of Data and Analytics Leaders

I had the chance to attend Cloud Native Security Con North America 2023, which ran Feb. 1 and 2 at the Seattle Convention Center.