DEV Community: unni mana

AI based Puzzle Book

unni mana — Tue, 09 Jun 2026 02:26:02 +0000

Rapid acquisition of knowledge often requires a hands-on approach, as true learning frequently occurs through direct application. To enhance one's understanding of Artificial Intelligence, it is essential to first master the fundamental terminology associated with the field. This foundational knowledge serves to re-orient the reader's perspective as they explore more complex concepts within the discipline.

Reflecting this pedagogical philosophy, I have developed a word search puzzle book dedicated to the subject of Artificial Intelligence. While the format mirrors traditional puzzle publications, the content is specifically curated to focus on AI-related themes. This interactive medium allows learners to familiarize themselves with complex nomenclature in a low-stress environment, reinforcing memory through visual recognition.

The book contains fifty specialized puzzles designed to stimulate cognitive engagement and strengthen the conceptual grasp of readers seeking to contribute meaningfully to this domain. Each puzzle is meticulously crafted to cover a diverse range of sub-topics, ensuring a broad exposure to the evolving landscape of AI technology and theory.

The volume is systematically organized to include comprehensive solutions for every puzzle, supplemented by a glossary that provides detailed explanations of key Artificial Intelligence terms. By combining the challenge of word discovery with a structured reference guide, the book serves as both a recreational activity and a robust educational tool for aspiring practitioners.

AI Word Search Puzzle

I Wrote a Book on Quarkus (And Here is a Free Chapter to Prove It)

unni mana — Fri, 22 May 2026 04:33:18 +0000

For years, Spring Boot was my go-to for Java microservices. It worked, so I didn't question it. But as we started moving more services to Kubernetes and AWS Lambda, the problems became impossible to ignore.

Cold starts taking 5-10 seconds.
Containers eating up 800MB+ of memory.
Wasting money just to keep services alive.

Then I found Quarkus. It starts in ~0.01 seconds as a native image and runs on a fraction of the memory. It was a game-changer.

So I wrote a book about it.

*📚 The Book: Practical Quarkus: From Zero to Native
*
It is a hands-on guide for Java developers who want to build cloud-native microservices that actually work. Every command and code block in the book has been tested on a real machine. No fluff. No copy-paste errors.

You can find it here on Amazon: Practical Quarkus: From Zero to Native

But I am not here to just ask you to buy a book. I am here to give you something useful first.

🎁 Your Free Chapter: "Why Quarkus?"

Since the first 10% of the book is free on Amazon as the "Look Inside" sample, I want to share exactly what you will learn in Chapter 1. It answers the most important question: "Should I even care about Quarkus?"

Here is a taste.

🚀 Build Your First Quarkus Endpoint (In Under 10 Minutes)

This is the exact code from Chapter 1. You can run it on your laptop right now.

Create a new Quarkus project using Maven:

mvn io.quarkus.platform:quarkus-maven-plugin:3.14.0:create \ -DprojectGroupId=org.example \ -DprojectArtifactId=code-with-quarkus \ -Dextensions=resteasy-reactive cd code-with-quarkus ./mvnw quarkus:dev

*2. Add a REST endpoint:
*
Create the file src/main/java/org/example/GreetingResource.java:

`package org.example;

import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.Produces;
import jakarta.ws.rs.core.MediaType;

@path("/hello")
public class GreetingResource {

@GET
@Produces(MediaType.TEXT_PLAIN)
public String hello() {
    return "Hello from Quarkus. Spring Boot took 3 seconds. I took 0.3.";
}

3. See it works:
Open your browser to http://localhost:8080/hello

You will see the message: Hello from Quarkus. Spring Boot took 3 seconds. I took 0.3.

That is it. You just ran your first Quarkus app. The dev mode even supports hot reload—change the message and save the file, and it updates instantly. No restart.

Use Quarkus if... Stick with Spring Boot if...
Startup time under 1 second matters: Startup time doesn't matter (long-running servers)
Running on Lambda, Fargate, or Knative: Running on traditional EC2 or dedicated servers
Memory is expensive (< 200MB per pod): You have plenty of RAM (2GB+ per pod)
You want native compilation with GraalVM: You rely on heavy reflection or dynamic proxies
Building 20+ microservices: Building one monolithic application

*📖 What's Next in the Book?
*
If you found this useful, the book goes much further. The first four chapters (available now) cover:

Chapter 2: Full CRUD with PostgreSQL and Panache (20 lines of code)

Chapter 3: Validation and proper HTTP error handling (404, 400)

Chapter 4: Logging and configuration for dev/prod environments

The book is 31% complete. When you buy it now, you pay once and get all future chapters (including the one on native executables) for free.

🔗 Get the Book & Code

Amazon (Live Now):[Practical Quarkus: From Zero to Native](https://www.amazon.com/dp/B0H2JR9DLD)

GitHub (All code examples): (github.com/unnivm/practical-quarkus)

*❓ Questions?
*
Ask me anything in the comments below. I will answer every one.

If you pick up the book, please leave a review on Amazon. It helps other Java developers find it.

Happy coding.

The Future of Autonomous Innovation: Inside the Gemini Enterprise Agent Platform|Google Cloud Next '26

unni mana — Sat, 25 Apr 2026 16:29:08 +0000

The landscape of artificial intelligence is shifting from static models to dynamic, autonomous entities. At Google Cloud Next '26, the unveiling of the Gemini Enterprise Agent Platform marked a pivotal moment in this evolution. Designed as a comprehensive ecosystem, the platform empowers organizations to build, scale, and manage production-ready AI agents capable of operating with a degree of independence previously confined to science fiction. To demonstrate this power, Google showcased a complex marathon simulation in Las Vegas, where hundreds of agents collaborated to manage everything from logistics to security in real-time.

Building the Foundation with ADK

At the heart of the developer experience is the Agent Development Kit (ADK). This toolkit simplifies the creation of modular agents by providing ready-to-use skills. A critical component of the ADK is its integration with the Model Context Protocol (MCP), which enables agents to seamlessly connect with Google Cloud services. This modularity ensures that developers don’t have to reinvent the wheel for every new capability; instead, they can assemble sophisticated agents that are deeply integrated with their existing cloud infrastructure from day one.

Universal Collaboration: A2A and the Agent Registry

One of the most significant hurdles in multi-agent systems is communication. The Gemini platform solves this through the Agent-to-Agent (A2A) protocol. This universal standard allows agents to advertise their specific capabilities and communicate with one another without the need for brittle, manual API integrations. Supporting this is the Agent Registry, a central directory functioning much like a DNS for AI. It allows agents to discover peers across a network, resolve identities, and find the specific skill sets required to complete a complex task collaboratively. Furthermore, the A2UI feature allows these agents to generate their own dynamic interfaces, ensuring they can interact not just with each other, but with human users in a friendly and intuitive manner.

Securing the Frontier: Red and Green Agents

Security in an autonomous world is paramount. The platform introduces a sophisticated "Red vs. Blue" dynamic powered by AI. The Red Agent acts as a "Friendly Pentester," an AI-powered security specialist that probes environments to identify exploitable risks. It doesn’t just scan for vulnerabilities; it validates them by analyzing attack paths—the actual route an intruder might take from the public internet to sensitive data. This provides a realistic view of runtime risks that traditional code analysis often misses.

The Fixer: Green Agent Integration

When a vulnerability is found, the Wiz-integrated Green Agent steps in. While the Red Agent finds the holes, the Green Agent is the "fixer." It suggests root-cause remediations, such as downgrading IAM privileges or patching authentication bypasses. It provides full transparency, showing developers the exact steps taken to discover the flaw. Most impressively, it can initiate developer workflows to apply these fixes directly to the code, closing the loop between threat detection and resolution.

Enterprise Management and Scalability

Beyond development and security, the platform excels in operational management. With Memory Bank and session management, agents remain stateful, recalling previous interactions and learnings. Specialized knowledge is provided via Retrieval-Augmented Generation (RAG) and AlloyDB vector functions, allowing agents to understand context like local city regulations. For operations teams, Cloud Assist provides observability, allowing for natural language debugging and proactive fixes. The infrastructure itself is built to scale, moving effortlessly from Cloud Run for simpler tasks to Google Kubernetes Engine (GKE) for massive, multi-agent simulations.

Accessibility and Open Innovation

Google is committed to making this technology accessible. The platform supports no-code integration, allowing teams to create agents using natural language prompts via Gemini Enterprise. To further foster innovation, Google has open-sourced the entire code for the Las Vegas marathon simulation. This repository and accompanying lab materials provide a roadmap for developers worldwide to start building the next generation of autonomous agents on a platform designed for safety, scale, and collaboration.

LLM Observability tool

unni mana — Sat, 25 Apr 2026 06:21:53 +0000

I just created an open source LLM observability named "TraceLM" that can track the following activities:

token cost
token length
latency information etc

The following tool can be found here:

https://github.com/unnivm/tracelm-java-agent

This tools offers a very cool dashboard to track the above information. The user is required to login to access this dashboard.

This tool currently supports only Java based applications as it is basically working based on instrumentation technique. But, this tool will be extended ti support other programming languages as well.

This project is under development and am going to add more features to to so that the idea is to make it an enterprise ready product.

A Leisurely Developer

unni mana — Mon, 30 Mar 2026 15:24:42 +0000

This is a submission for the 2026 WeCoded Challenge: Echoes of Experience