# xprilion's blog

> Software development, cloud computing, AI/ML, and developer tools by Anubhav Singh.

## Blog Posts

- [74% better Hindi for IBM Granite 4.1 on a Laptop GPU](https://xprilion.com/teching-ibm-granite-hindi-on-a-laptop-gpu.md): The model beats Opus 4.6 on table extraction, a sweet use-case for an SLM, and supports 10+ languages, alas, no Hindi. I decided to change that.
- [Autoscaling LLM Inference on GKE with TPU v5e and vLLM](https://xprilion.com/autoscaling-llm-inference-on-gke-with-tpu-v5e-and-vllm.md): This post documents the end-to-end process of deploying an autoscaling LLM inference endpoint on Google Kubernetes Engine using Cloud TPU v5e and vLLM. It covers everything from quota discovery to a l
- [Building YouTube Shopping AI with Gemini 3 Flash](https://xprilion.com/youtube-shopping-ai-with-gemini-3-flash.md): YouTube-first shopping guides with Next.js + Gemini + Programmable Search
- [Golang vs Rust: API Performance Load Test](https://xprilion.com/golang-vs-rust-api-performance-load-test.md): Roughly 8 months back, I wrote [an article comparing the performance of Python against Golang](https://xprilion.com/golang-vs-python-api-performance-load-test/), and the latter emerged a clear winner.
- [What's Next for Call Centers? Embracing AI in 2024](https://xprilion.com/whats-next-for-call-centers-embracing-ai-in-2024.md): Stuck at a problem with your newly bought printer? Or maybe your order is being shipped to another country without your consent? Like, you turned to a call-based customer support session to get it sol
- [Ollama Models on Cloud Run](https://xprilion.com/ollama-models-on-google-cloud-run.md): This blog is a read-along for the repository [xprilion/ollama-cloud-run](https://github.com/xprilion/ollama-cloud-run/) which shows how to deploy various models using the Ollama API on Cloud Run, to r
- [Firebase Data Connect - Getting Started](https://xprilion.com/firebase-data-connect-getting-started.md): tl;dr - Demo: [https://dataconnectdemo-x.web.app/](https://dataconnectdemo-x.web.app/), Github: [https://github.com/xprilion/firebase-data-connect-demo](https://github.com/xprilion/firebase-data-conne
- [How much Randomness do you really need?](https://xprilion.com/how-much-randomness-do-you-really-need.md): > tl;dr: This article doesn't tell you how to generate more random numbers, it asks you - do you need it? And also sort of tells you how to generate more random numbers along the way.  In a world full
- [Resume Template by Anubhav - Beat the ATS](https://xprilion.com/resume-template-by-anubhav.md): tl;dr - I made [a resume template that scores high on ATS systems and has become insanely popular](https://www.overleaf.com/latex/templates/resume-template-by-anubhav/dhmkrwtksdgy)! Since then, I have
- [Golang vs Python: API Performance Load Test](https://xprilion.com/golang-vs-python-api-performance-load-test.md): In 2016, when I had started learning Python, I was increasingly convinced it would be the last programming language I would learn. After countless projects created with Python, a couple of technical b
- [Build a Gemini Store with Google Sheets](https://xprilion.com/build-a-google-gemini-pro-store-with-google-sheets.md): Giving personalities to your Gemini APIs via Gemini Store in Google Sheets
- [Revamping my blog with Hashnode Headless CMS](https://xprilion.com/gatsby-hashnode-headless-starter-kit.md): This blog has been written for the Hashnode Headless API Hackathon and the Gatsby + Hashnode Headless + Typescript is publicly available at - [https://github.com/xprilion/gatsby-hashnode-headless](htt
- [The Story of the Wright Brothers and Samuel Pierpont Langley](https://xprilion.com/the-story-of-the-wright-brothers-and-samuel-pierpont-langley.md): *In the dawn of the 20th century, the quest to conquer the skies was akin to a modern space race. The central figures in this drama were the Wright brothers, Orville and Wilbur, and a distinguished fi
- [Docker Python Dump HTTP Request Example](https://xprilion.com/docker-python-dump-http-request.md): This blog is an explanation for the code present in the [github.com/xprilion/docker-python-dump-http-request-example](http://github.com/xprilion/docker-python-dump-http-request-example) repo.  The mai
- [Job Scheduling on Google Cloud Platform](https://xprilion.com/job-scheduling-on-google-cloud-platform.md): > Job scheduling is like conducting an orchestra; every task must play its part at the right time to create a harmonious symphony of efficiency.  In the world of software systems, job scheduling plays
- [Python Websockets SSL with Let's Encrypt](https://xprilion.com/python-websockets-ssl-with-lets-encrypt.md): This tutorial is an explanation of my gist [Python Websockets SSL with Let's Encrypt](https://gist.github.com/xprilion/ceab48ec77a70be1d403e396170991e6) .  With the launch of HTML5 in 2008, a technolo
- [High Availability ML Deployments](https://xprilion.com/high-availability-ml-deployments.md): > The average cost of IT downtime is $5,600 per minute.  ~ Gartner  Downtimes can be costly. During downtimes, a company may face loss of business, loss of customer trust, loss of reputation in the te
- [Dividing is Easy, Unite if you can](https://xprilion.com/divide-vs-unite.md): There’s always an easy option and one that needs courage, patience and a strong will.  When I was a kid, every evening, I used to go out to play in a park nearby my home. We were a good mix of kids
- [The Cost of Standing Out](https://xprilion.com/cost-of-standing-out.md): 9 years back, on this date, I launched a social network built from scratch. Almost a month later, I added on the same platform a search engine, again, built from scratch.  The website gained almost 60
- [Decoding Propaganda In Memes](https://xprilion.com/decoding-propaganda-in-memes.md): Memes - the nice images you see floating around literally everywhere on the Internet these days and the stuff that fuels your social media. Memes began as a source of entertainment - containing jokes
- [Working with Amazon Keyspaces Cassandra distribution using Python](https://xprilion.com/python-amazon-keyspaces-cassandra.md): [Cassandra](https://cassandra.apache.org/doc/latest/) is a popular NoSQL database with capabilities to handle massive data by using a distributed array of commodity hardware. After a boring introducti
- [A Generic MongoDB Wrapper API with Flask and PyMongo](https://xprilion.com/generic-mongodb-wrapper-using-flask-pymongo.md): ---  Hey there, hope you're well! It's been a while since I wrote something here, to my defence I wrote this tutorial about [How to setup a secure, remote JupyterLab workstation](https://www.digitaloc
- [Machine Learning On 2GB RAM](https://xprilion.com/ml-on-2gb-ram.md): If you think you read the title, or maybe I typed it wrong, you're wrong in both cases. The title proposes an article about performing machine learning with the bare minimum RAM usage, and that's what
- [Methods of Integrating Artificial Intelligence on Flutter](https://xprilion.com/methods-of-integrating-artificial-intelligence-on-flutter.md): ---  One of the coolest UI toolkit's for cross-platform applications, [Flutter](https://flutter.dev), a Google offering has grown rapidly in popularity over the last few years. While there are specula
- [Folding@Home Setup - Local and Cloud VM](https://xprilion.com/folding-at-home-setup.md): ---  The world is facing an unprecedented situation with the [Coronavirus outbreak](https://www.who.int/emergencies/diseases/novel-coronavirus-2019). It is crucial at these times for everyone to join
- [Being Noob - Starting a tech journey](https://xprilion.com/being-noob.md): Hi there! I realized I began sort of suddenly, talking about networks first and then a tech scholarship, while not having yet talked about how things began for me and what I believe is a good way to l
- [What is the Venkat Panchapakesan Memorial Scholarship and Why its not won in a day](https://xprilion.com/venkat-panchapakesan-memorial-scholarship-part-1.md): It was with a crazy smile and immense gratitude that I read through the entire mail from Google India about being awarded the [Venkat Panchapakesan Memorial Scholarship](https://buildyourfuture.withgo
- [Difference between localhost, 127.0.0.1 and 0.0.0.0](https://xprilion.com/difference-between-localhost-127001-and-0000.md): I've been often asked, why do we sometimes need to specify the IP for a program to be `0.0.0.0` in order to get it to be accessible from machines outsides the network. I shall attempt to answer the qu
- [OMG, a new look!](https://xprilion.com/hello-world.md): Hey there! If you have visited my website in 2019 or before, you would know that my website was pretty blue and very blue. Overall, it was unexciting, and there was no fresh content on it. But did tha

## Codelabs

- [Hello, WebMCP](https://xprilion.com/codelabs/hello-webmcp.md): A noob's walkthrough to your first WebMCP implementation
- [Transforming MCP Servers to Gemini CLI Extensions](https://xprilion.com/codelabs/transforming-mcp-servers-to-gemini-cli-extensions.md): Convert your existing MCP Servers to Extensions for Gemini CLI
- [Google ADK Voice AI Agent](https://xprilion.com/codelabs/google-adk-voice-ai-agent-wandb-weave.md): Let's build a Bidi streaming voice AI agent with Google ADK
- [Quick Gemini in React apps using Firebase Vertex AI](https://xprilion.com/codelabs/quick-gemini-in-react-apps-using-firebase-vertex-ai.md): Learn how to quickly integrate Gemini into your React Apps using Firebase Vertex AI
- [Build a Gemini Store with Google Sheets](https://xprilion.com/codelabs/build-a-google-gemini-pro-store-with-google-sheets.md): Giving personalities to your Gemini APIs via Gemini Store in Google Sheets
- [Google Cloud Run example using Python Flask](https://xprilion.com/codelabs/google-cloud-run-python-flask-example-lab.md): An example of deploying a Python Flask application to Google Cloud Run.
- [Python Flask app with Gemini Pro Vision Multimodal](https://xprilion.com/codelabs/gemini-pro-vision-multimodal-python-flask-generative-ai-example-lab.md): A starter example of using Gemini AI Pro Vision Multimodal model via Google AI Studio (formerly MakerSuite).
- [Introduction to Building Solutions with PaLM2 API via VertexAI on GCP](https://xprilion.com/codelabs/vertexai-palm-api-introduction-python-lab.md): A starter guide to building innovative solutions with PaLM2 API via Vertex AI on GCP.

## Pages

- [Git Shortcuts](https://xprilion.com/gitshort.md): Some git shortcuts that live permanently on my machines.
- [Meet](https://xprilion.com/meet.md): Looking to get on a Google Meet with me?   Click here - [https://meet.google.com/rxj-gzjp-dhb](https://meet.google.com/rxj-gzjp-dhb)  Or copy paste the code below on your Google Meet   ```bash rxj-gzj
- [Say My Name](https://xprilion.com/say-my-name.md): Names - so many of them! Here's a little guide on how to pronounce mine right -   Anubhav Singh  अनुभव सिंह  **/ʌ.nʊ.bʌʋ sɪŋɡʱ/**       Play      ### Detailed Breakdown: - **
- [Tools](https://xprilion.com/tools.md): Over the years, I have built a bunch of tools which help me on an everyday basis. Here are some which you can have a go at: - [**LinkBorg**](https://linkb.org/) *Purpose:* Link shortener and single pa
- [Talks](https://xprilion.com/talks.md): I enjoy talking to people about what I have learnt from my experimentations with technology. Sometimes I simply talk about my experiences and sometimes, I help them learn something new. Here are a few
- [Publications](https://xprilion.com/publications.md): Writing has always been a strong passion within me. Over the years I have written blogs, pieces and books, in a hope that someday they'll help some explorers on their journey towards seeking knowledge