ANDMORE

Software Engineering Manager

Levatas

Development Team Lead

Senior Software Engineer

EndToEnd Management

Senior Software Engineering Consultant

Laura Software

Senior Software Developer

Independent 

Independent Software Developer

Software Engineer Manager at Juniper By IMC

Software Engineer Manager with over 15 years of experience in software development and three years of experience in management. Proficient in various programming languages, including [C#, Node, Javascript, and Python]. 

Experienced in leading cross-functional teams to deliver complex software projects on time and within budget. 

Strong background in agile development methodologies and project management.

Extensive experience in the entire life cycle of the software design process, including requirements definition, prototyping, design, implementation, testing, and maintenance.

This single book changed forever the way I build systems.


Kleppmann mixes theory with real-world examples, so it's not just abstract concepts – it's stuff you can use in your day-to-day work.

The book covers a lot:

- Transactions
- How to scale systems
- Keep systems reliable
- Make systems efficient while dealing with tons of data.


You can learn and apply new concepts to solve practical problems in data-intensive applications.

I found myself coming back and re-read it from time to time.

What would you recommend after re-reading this?

Raúl Junco

Tiger Abrodi

Publisher + Subscribers = Observer pattern


If you already know my newsletter, you know how this pattern works. (Link in the comments)

A publisher (subject) and its subscribers is a good way to visualize this pattern.

The Observer Pattern defines a one-to-many relationship between a set of objects. When the state of one object changes, all of its dependents are notified.


The real power of the observer lies in loosely coupled components. They can interact but typically have very little knowledge of each other.

The subject knows only that an observer implements a certain interface (the Observer interface). 

It doesn't need to know the observer's concrete class, what it does, or anything else.


We can add new observers anytime because the subject depends only on a list of objects that implement the Observer interface.

We can replace any observer at runtime with another observer, and the subject will keep sending notifications.


We can also remove observers at any time.

We never need to change the subject to add new types of observers.

Let's say we have a new concrete class that needs to be an observer. We don't need to change the subject to accommodate the new class type.

The subject doesn't care; it will notify any object that implements the Observer interface.

Changes to either the subject or an observer will not affect the other.


Loosely coupled designs allow you to build flexible systems because they cut the interdependency between objects.


Stay decoupled and Subscribed!

Dennis Appiah

Queues are at the core of modern Software.

A simple introduction to Queues.


At the heart of queues lies the First-In-First-Out (FIFO) principle.

Similar to people waiting at a ticket counter — the first person in line is the first to get service.

• Enqueue: Add an element to the back.
• Dequeue: Remove the front element


Linear Queue

A basic queue with a fixed size.

Once full, it can't accept new elements.

It follows a simple structure akin to a daily traditional queue.

Example: a print job queue where documents wait in line to be printed in the requested order.


Circular Queue

This variant of the linear queue wraps around, allowing elements to be added and removed in a circular manner.

When the queue reaches its maximum size, it starts overwriting the oldest elements.

Used by embedded systems and device drivers.


Priority Queue 

This queue assigns priorities to its elements.

It dequeues the element with the highest priority first.

It is ideal for scenarios requiring task prioritization and resource management.


Real-World Applications:

• Task Scheduling: Manage pending tasks, processing them in the order received.

• Background Processing: Handle long-running tasks asynchronously.

• Message Brokers: It facilitates communication in distributed systems or microservices.


TL;DR

Enqueue to add, dequeue to remove—simple yet powerful!

Queues are the backbone of modern software architecture, providing a structured and efficient way to:

• Manage data flow
• Coordinate tasks
• Handle concurrency


Knowing about queues will help you become a better SWE. Master them!

Michell Campos

The only time you shouldn't run Performance tests is if you don't care how your API works in production.

4 types of performance tests and what they tell you. 


The idea is to understand how your API will perform under real-world conditions.


1. Load Tests

This test checks the API under an expected load, the usual number of users or transactions it needs to handle. 

You will learn how many users or transactions it can support without degradation.

Load tests help you understand everyday operational performance.


2. Stress Tests

This test determines the limits of your API by increasing the load until it becomes unstable. 

This helps you understand how much your application can handle before it fails. But most importantly, how does your API recover from these failures?

Stress tests help you prepare for extreme conditions.


3. Spike Tests

This test examines how well the API handles sudden traffic increases.

If you are running any promotion, you should be using these tests.

Spike tests ensure your API can handle sudden bursts of traffic and can scale without problems.


4. Soak Tests

The idea is to check the API's performance under an average load extended over a long period.

It will show you issues that might not appear in shorter tests, like:
- memory leaks 
- slowdowns that develop over time
- resource depletion

Soak tests uncover long-term stability and performance issues.


Is your API delivering what your users expect?

Gurmukh Singh

If you have been in the cloud for more than 15 minutes, you know how difficult it is to test and debug serverless applications (Lambdas, SQS, etc.)

 
What if I showed you how to debug your cloud applications locally with fast iterations? 

This is why Winglang caught my attention.

Wing is a framework designed to develop cloud applications.

The core concept is that you can specify resources directly within your application.

What shocked me was that I didn't need to leave my local environment to create resources or wait for the cloud to provision them.

I can run a function locally that listens to a queue and saves values to a bucket.

Using the Winglang console, I can run a local simulation and visualize what happens in each step.

It feels like debugging.


𝗪𝗶𝘁𝗵 𝗪𝗶𝗻𝗴𝗹𝗮𝗻𝗴:

- You Code
- You Test locally
- You Compile it
- You Deploy to a cloud provider


𝗧𝗵𝗲𝗿𝗲 𝗮𝗿𝗲 𝘀𝗼𝗺𝗲 𝗼𝘁𝗵𝗲𝗿 𝗮𝗱𝗱𝗶𝘁𝗶𝗼𝗻𝗮𝗹 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀:

- Supports Hot Reloading, to get instant feedback (I’m talking milliseconds).

- It creates the necessary security policies (this alone saved me hours).

- It generates CloudFormation / Terraform.

- There is no need to learn the syntax for each cloud provider. Your code can be compiled to AWS, GCP, Azure, or any custom platform.


The community is great, and the framework currently supports these outputs:

- AWS CDK platform
- Terraform/AWS platform
- Terraform/GCP platform
- Terraform/Azure platform
- Simulator platform
- Custom platforms


Open-source is taking over the world, and Wing is part of it. 

Give them a star: https://git.new/winglang


Big thanks to the team at Wing for collaborating with me on this post.

Only 13% of data projects make it into Production.


Data is not enough; you need: 

1. A pipeline that connects to your data sources. 
2. To train your models.
3. Build a UI fast enough to render the result and chart with millions of records.
4. Finally, deploy your product so your customers can use it.


Most people try to build this from scratch. Instead, I use Taipy.

- You can create pipelines connected to your data sources (Databricks, Snowflake).

- You can transform your data. 

- You can train and test your models. 

- You can connect my pipeline results to the Frontend.

- Taipy is cloud agnostic. You can deploy on any cloud platform regardless of where the data resides and how it is consumed.

- You only need to know the tech stack: Python.

- Taipy is open source; I can contribute if I need. 


I’m obsessed with putting things in front of clients as soon as possible.

I use Taipy because I can spin a prototype with a few lines of code and put a product in front of my users.


Take a look at the examples in their repo and give a star to this cool library: https://lnkd.in/eYsaZGEr


Big thanks to Taipy for collaborating with me in this post.

I will build my next Dashboard with 0 lines of UI code.


Building a UI for your back end takes days and a lot of code.

This is exactly what Taipy has fixed with its new tool, Taipy Designer.

You can drag and drop UI widgets onto a canvas and connect them with your variables.

When you open Taipy Designer with your code, all these variables are automatically accessible.


𝟱 𝗦𝗶𝗺𝗽𝗹𝗲 𝘀𝘁𝗲𝗽𝘀 𝘁𝗼 𝗯𝘂𝗶𝗹𝗱 𝗮 𝗗𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 𝗾𝘂𝗶𝗰𝗸𝗹𝘆:

1- Drag and drop a bar chart widget onto the canvas. 

2- Select a data source from your list of variables.  

3- Resize and customize the widget with your own styles. 

4- Then add a ‘Date Selector’ widget connected to your local date variable, allowing users to filter the data as needed.

5- From here, you can add more UI elements, a map, a switch, and many tabs.


The cool part is that you can build all this without adding a single line of code. Taipy will connect this UI with your backend code.

Check this video and watch how fast you can build a ready-to-production data application.


If you hit the market first, you will have the edge; that’s why Taipy makes sense: it will save you time and money.

Give a star to this cool library here: https://lnkd.in/eYsaZGEr

Thanks to Taipy for sharing this video and collaborating with me on this post.

Taplio

Raul Junco's Linkedin Analytics

Raul Junco

What is Raul talking about?

Raul Junco's Best Posts (last 30 days)

Want to drive more opportunities from LinkedIn?