Applying for jobs are fun, but it is not fun to write cover letters at least not compare to hacking on some new open source tool or testing out something new. Therefore the goal is to reduce the time to apply for a job thus the goal today is to generate cover letters. For this we will need the following: CV, I will copy paste mine from linkedin(minimal effort and I spend way to much time on linkedin so it is up to date) Job add(will take one from linkedin that sounds interesting) Get the CV.…

This blog post is a collection of awesome go resource and will be continuously update. The goal is that each resource should be describe shortly as well. Blogs: Generics can make your Go code slower Effective go Go memory model inline Videos: Obscure Go Optimisations A Guide to the Go Garbage Collector …

In this blog post, we will discuss how to store Kubeflow Pipeline templates in GCP Artifact Registry, enabling reusability and version control for your pipelines. Using Artifact Registry over Cloud Storage simplifies version control and allows for easier collaboration between single or multiple users. The Kubeflow Pipelines SDK registry client is a new client interface that you can use with a compatible registry server (ensure you are using the correct KFP version), such as Artifact Registry, for version control of your Kubeflow Pipelines (KFP) templates.…

This blog post will go over how to build and run your very first kubeflow pipeline (kfp). In short, Kubeflow Pipelines is a platform for building and deploying portable, scalable machine learning (ML) workflows based on Docker containers. There are a lot of possibilities to run the pipelines, but in this series, we will use gcp vertex pipelines. Vertex will be the runner, but the pipelines will follow the kubeflow conventions meaning you can run them on whatever platform at hand or host kubeflow on your own Kubernetes cluster.…

This blog post provides a step-by-step guide for setting up your VS Code dev container to work with Google Cloud Platform (GCP) services and APIs by configuring default GCP credentials. By authenticating your application with your GCP credentials, you can access the necessary resources without requiring additional authentication steps, saving time and streamlining your development workflow. Configuring Default GCP Credentials To use default credentials with GCP, you can follow the steps below:…

VS docker Dev Container Are you tired of dealing with messy, inconsistent development environments that slow down your workflow? Look no further than VS Code dev containers, a powerful feature in Visual Studio Code that can streamline your development process and improve consistency across different environments. In this post, we’ll dive into what dev containers are and how you can use them to supercharge your development workflow. We’ll also walk through how to set up a dev container for a Python environment and you will get an understsnding how easy it is to get started.…

This blog post will dive into building postgres extensions using pgx in rust. In order to do something existing we will ride the hype curve and integrate ChatGPT3 into postgres. TLDR: repo Setup First step is to set up pgx: $ cargo install --locked cargo-pgx $ cargo pgx init We will then create a new create using: $ cargo pgx new my_extension $ cd my_extension Which should give you something like:…

As part of the go 1.20 release memory areas where introduced to the standard lib but not mentioned in the release notes but is still being discussed as of 2023-03-02 here. Memory arenas allow users to allocate memory and are described by the docs as: The arena package provides the ability to allocate memory for a collection of Go values and free that space manually all at once, safely. The purpose of this functionality is to improve efficiency: manually freeing memory before a garbage collection delays that cycle.…

This blog post will go in to depth how to use shared memory together with nvidia triton and pinned memory for model serving. This will continue to build further on the other blog posts related to triton. First we will focuse on shared memory and then move over to also look in to pinned memory and why it matters. Shared memory In the triton examples(python) shared memory is often abbreviated as shm.…

Current hack to just get the stuff working.…