Owen Colegrove’s Post

Just obtained a new badge: Inspect Rich Documents with Gemini Multimodality and Multimodal RAG Skill Badge from Google Cloud Skills Boost! During this extensive training I learned and practiced concepts like RAG (Retrieval Augmented Generation) and Google's DocumentAI API. This badge is rewarded for completing a Challenge Lab that tested what I have learned from the course. I really like this approach by Google. Instead of a short MCQ test at the end of a course for certification, this is a really great way to really test whether the user really learned what he was supposed to learn. #GoogleCloudReady

Learn about Graal Cloud Native now being Graal Development Kit for Micronaut, as well as the general availability of Graal Development Kit for Micronaut 4.3.7 based on Micronaut® framework 4.3.7. https://lnkd.in/g7h-3vk9

The fundamentals often trip me up, so when legend himself Darko Mesaroš said let's make some videos, I was all in! Here is the first of a series of videos for code fundamentals with a cloud twist, starting with how building robust error handling can turn a fragile application into an unshakeable fortress! https://lnkd.in/erg6T3AU #PythonProgramming #ErrorHandling #CodeTutorial

Leptos + WASI 0.2 for full stack #Wasm development! Great post from Enzo Nocera and Brooks Townsend on how the wasi:http standard made integration into popular #Rust library #Leptos possible + how, thanks to wasmCloud's emphasis on standards, devs can now run it on Cloud Native Computing Foundation (CNCF) wasmCloud! ⚡ https://lnkd.in/g4BwxQJy

Learn about Graal Cloud Native now being Graal Development Kit for Micronaut, as well as the general availability of Graal Development Kit for Micronaut 4.3.7 based on Micronaut® framework 4.3.7. https://lnkd.in/gw3VTecP

Learn about Graal Cloud Native now being Graal Development Kit for Micronaut, as well as the general availability of Graal Development Kit for Micronaut 4.3.7 based on Micronaut® framework 4.3.7. https://lnkd.in/dxPJ24Dz

I created this pricing comparison table this week for LLMs served as PaaS by the token and some interesting things reveal themselves here. #AWS and #GCP are clearly going for model diversity while both AI21 Labs and Anthropic are making sure that prices are the same across clouds. #Azure is mostly sticking with their investment in OpenAI although you can run Mistral, Llama, and Cohere inside functions so the pricing structure isn't exactly equivalent. I pulled all this from public pricing pages, but what did I miss here? What other patterns do you see in PaaS LLM hosting? cc: Brent Blawat Nathan Cartwright Anthony Placeres Jeff Myers Joe Markwith Roger Campbell Casey Bleeker John Wynkoop Tige Phillips

Deploying an ML application becomes a lot easier with the help of Docker. TBH I didn't expect it will be that much smooth, involving just 3 simple steps: 1. Develop and Dockerize an ML application and push the image to Docker hub. 2. Pull the image from Docker hub to Google Cloud Container registry using Google Cloud SDK Shell. (Note : you can directly push the image to Container Registry without involving Docker hub ). 3. Once pulled, go to Google Cloud Run and deploy the ML Application with just few clicks. Here I've developed an Machine Learning Application which will predict whether a customer will make purchase or not depending on their age, previous purchase history and number of pages visited. The application is capable of making online prediction as well as batch prediction (for a list of customers). The code for the application can be fount here : https://lnkd.in/dZVWAVvq #Docker #GCP #MachineLearning #Deployment #CloudRun #ContainerRegistry

ETCD : Distributed Key value store for the distributed system. It is used in kubernetes. Distributed in the definition is most important. It uses Raft algorithm for Consensus Protocol for leader election and log replication. Logs should be consistent among leaders and followers. This gives Fault tolerance, scalability ,consistency, Load balancing. Application: It stores events of Kubernetes component and only api server can talk to it. ETCD work like a state machine ,,it looks one from outside but it is distributed by nature. Application of etcd via k8 are - container orchestration,, edge computing,cloud like EKS. #etcd #Kubermates #cloud #Algorithms #distributedSystem

Been working on PyTorch but always wanted to pick up JAX? Google Cloud just dropped "The PyTorch developer's guide to JAX fundamentals". Contains a side-by-side implementation of a training loop in both Pytorch and JAX + Flax NNX for those interested in exploring the JAX world in familiar terms. Here’s what it offered: ✨ Side-by-side comparisons of training loops in PyTorch vs. JAX + Flax NNX. ⚡ Learn how JAX’s speed, scalability, and vectorized computation can transform your projects. 🎯 Perfect for anyone curious about harnessing JAX’s cutting-edge capabilities without starting from scratch. This guide makes the transition easy, practical, and downright exciting. Whether you’re looking to enhance your workflows or just stay ahead in the ML landscape, this is your next move. Link: https://lnkd.in/g-ZFFmSV #ArtificialIntelligence #DeepLearning #DataScience #PyTorchToJAX