Amazon launched a new AI agent

Welcome back!

We’ve been solving easy problems lately. Let’s try a medium difficulty problem today. Are you ready?

Today we will cover:

• Binary Tree Zigzag Level Order Traversal

• Pub-Sub vs Message Queues

Read time: under 4 minutes

CODING CHALLENGE

Binary Tree Zigzag Level Order Traversal

Given the root of a binary tree, return the zigzag level order traversal of its nodes' values. (i.e., from left to right, then right to left for the next level and alternate between).

Example 1:

Input: root = [3,9,20,null,null,15,7]
Output: [[3],[20,9],[15,7]]

Example 2:

Input: root = [1]
Output: [[1]]

Solve the problem here before reading the solution.

PRESENTED BY DEEL

Optimize global IT operations with our World at Work Guide

Explore this ready-to-go guide to support your IT operations in 130+ countries. Discover how:

• Standardizing global IT operations enhances efficiency and reduces overhead

• Ensuring compliance with local IT legislation to safeguard your operations

• Integrating Deel IT with EOR, global payroll, and contractor management optimizes your tech stack

Leverage Deel IT to manage your global operations with ease.

Download free guide

SOLUTION

To solve this problem, we'll use a modified level-order traversal using a queue. We'll keep track of the current level and reverse the order of nodes alternatively.

For each level, we'll process all nodes in the current_level, collect their values, and add their children to the queue for the next level. If the level is odd, we'll reverse the collected values before adding them to the result.

The time complexity of this solution is O(n), where n is the number of nodes in the binary tree.

SYSTEM DESIGN

Pub-Sub vs Message Queues

When building event-driven systems, you need a way to handle communication between different services. Two popular approaches are Message Queues and Pub-Sub systems. Let's understand when to use each one.

Message Queues work like a line at a coffee shop - messages wait in a queue and are processed one by one. When a service (producer) sends a message, it goes to the back of the queue. Another service (consumer) picks up messages from the front of the queue and processes them. Once a message is processed, it's removed from the queue.

In Pub-Sub systems, messages are broadcast to all interested services. Services that want to receive messages (subscribers) tell the system what types of messages they want. When a service (publisher) sends a message, all interested subscribers receive a copy of that message.

Message Queues are great when you want each message to be processed exactly once. For example, in an e-commerce system, when a customer places an order, you want exactly one service to process that order. You don't want multiple services trying to process the same order.

Pub-Sub systems shine when multiple services need to react to the same event. For instance, when a user creates a new account, you might want one service to send a welcome email, another to update analytics, and a third to check for fraud - all based on the same "new user" event.

Both approaches can be combined in modern systems. Tools like Apache Kafka and RabbitMQ support both patterns, letting you choose the right approach for each use case.

Here's a comparison of Message Queues vs Pub-Sub:

FEATURED COURSES

5 Courses of the Week

✅ Google Cloud Developer Certificate: Teaches cloud-native application design and maintenance through hands-on Google Cloud labs. Provides real-world experience with Google Cloud technologies for the Cloud Developer role.

✅ Microsoft Azure Developer Certificate: Covers end-to-end Azure solution development incl. compute, storage, authentication, and security. Provides interactive exercises using Microsoft Learn Sandbox for practical experience.

✅ IBM Back-End Development Certificate: Develops server-side skills in Linux, Git, Python, SQL, Django, Docker, and Kubernetes for entry-level positions. Includes multiple hands-on projects to build a professional portfolio. Bonus: Provides access to IBM's Talent Network for job opportunities upon completion.

✅ Standord’s Algorithms Specialization: Introduces algorithm fundamentals with emphasis on conceptual understanding rather than low-level implementation. Includes weekly quizzes and programming assignments in your language of choice. Prepares students for technical interviews and professional algorithm discussions.

✅ PHP Full Stack Bootcamp: Combines front-end (HTML, CSS, JavaScript) and back-end (PHP, MySQL, WordPress) development in one comprehensive course. Includes 80+ PHP projects and complete shopping cart website development. Prepares for freelance work and junior developer positions with certification.

WHAT'S NEW IN AI? 

Start learning AI in 2025

Keeping up with AI is hard – we get it!

That’s why over 1M professionals read Superhuman AI to stay ahead.

• Get daily AI news, tools, and tutorials

• Learn new AI skills you can use at work in 3 mins a day

• Become 10X more productive

Sign up and start mastering AI

NEWS

This Week in the Tech World

Amazon unveils Nova Act: Amazon launched Nova Act, an AI agent that can control web browsers and perform simple tasks independently. The agent will power features in the upcoming Alexa+ upgrade and outperforms similar tools from OpenAI and Anthropic on internal tests.

MrBeast advises Zuckerberg: YouTube star MrBeast told Mark Zuckerberg that Meta's lack of multilingual dubbing hurts content reach compared to YouTube. Zuckerberg agreed and suggested AI could solve the problem. MrBeast also noted that AI-generated dubs in his voice improve viewer retention.

Indian startups raise $2.5B: India's tech startups raised $2.5 billion in Q1 2025, up 13.64% from the previous quarter and 8.7% from last year. Auto tech, enterprise applications, and retail were the top-performing sectors, making India the third most-funded country globally.

23andMe files for bankruptcy: Gene testing firm 23andMe has filed for Chapter 11 bankruptcy protection to facilitate its sale. CEO Anne Wojcicki has resigned but will remain on the board. Despite going public in 2021 with a $6 billion valuation, the company never made a profit.

OpenAI launches image generator: OpenAI has enhanced ChatGPT with technology that can generate images from complex instructions, including four-panel comic strips. Unlike previous versions, the new system blends various concepts and draws on everything ChatGPT has learned from the internet.

Alibaba releases new AI model: Alibaba Cloud launched "Qwen2.5-Omni-7B," a multimodal open-source AI model that processes text, images, audio, and videos while generating real-time responses. The model aims to deliver "cost-effective AI agents" and can be deployed on edge devices like phones.

Anthropic reveals LLM insights: Anthropic researchers published two papers on how large language models process information. The findings show Claude has language-agnostic components, can plan ahead, adapt to changes, and performs multi-step reasoning rather than just memorizing training data.

ChatGPT image feature overloads servers: OpenAI CEO Sam Altman announced that viral use of ChatGPT's new image-generation AI is "melting" the company's GPUs. OpenAI will temporarily limit the feature's usage as they work to make it more efficient.

BONUS

Just for laughs 😏

HELP US

👋 Hi there! We are on a mission to provide as much value as possible for free. If you want this newsletter to remain free, please help us grow by referring your friends:

📌 Share your referral link on LinkedIn or directly with your friends.
📌 Check your referrals status here.

YOUR FEEDBACK

Until next time, take care! 🚀

Cheers,

Sahil + Sarra