Async vs Multithreading: Understanding Concurrency in Python

FastAPI is quickly becoming the default choice I see for new Python backend projects. What stands out to me is the combination of strong performance and a low barrier to entry. The framework’s async-first approach makes it easier to reason about concurrency early, and the built-in request validation encourages better API contracts from day one. For someone focused on backend development, that structure matters as much as speed. I also like that it nudges you toward clearer separation between routing, schemas, and database logic, which maps well to how real services are maintained. When you’re choosing a backend framework, do you optimise first for performance or for developer experience? #python #backend

Migrated from Node.js (Express) to FastAPI Recently, I rebuilt one of my APIs from Node.js to FastAPI — and here’s what changed: 🔹 Manual validation → ✅ Pydantic models 🔹 Custom Swagger setup → ✅ Auto-generated docs 🔹 More boilerplate → ✅ Cleaner & structured code 🔹 Middleware logic → ✅ Dependency Injection ⚡ Performance: FastAPI felt faster and more structured (especially with Python typing). 🛠 Maintenance: Easier due to built-in validation & documentation. If you're working with Python or ML stack, FastAPI is a solid choice for modern APIs. #Backend #FastAPI #NodeJS #Python #APIDevelopment

When I first saw our production database hitting 99% utilization, I thought we needed a bigger server. I was wrong. The real problem wasn’t infrastructure. It was one unoptimized query quietly running inside a loop. Development was fast. Production exposed everything. I just published a deep dive on Django Query Optimization — covering N+1 problems, select_related vs prefetch_related, aggregations, indexing strategies, and real-world mistakes that slow apps down. here is my article link https://lnkd.in/gViDwwVd If you’re building APIs in Django, this might save you from a painful production incident. Would love to know: Have you ever faced a performance issue that wasn’t obvious at first? #Django #BackendDevelopment #Python #DatabaseOptimization #WebPerformance

n8n is a powerhouse for workflow automation, offering the perfect balance between a user-friendly UI and the flexibility to inject custom logic where it’s needed most. Self-hosting n8n gives you incredible control, but things get interesting the moment you try to bring Python into the mix. I recently spent some time navigating the complexities of adding a Python runner to a self-hosted setup. While it sounds straightforward, the reality of Docker permissions and environment pathing can lead to a fair share of troubleshooting. I’ve documented the exact roadblocks I hit and the final fix that actually works. If you’re looking to supercharge your n8n workflows with custom Python scripts without the headache, this is for you. Check out the full guide here: https://lnkd.in/gwK4bvjH #n8n #Python #SelfHosting #Automation #DevOps #LowCode #DataEngineering

Are you ready to level up your backend development skills? I just released a comprehensive 4-hour crash course on FastAPI, one of the fastest-growing and most modern web frameworks for Python today. Whether you’re a Django/Flask fan or coming from a Node.js background like me, this video will show you why FastAPI is a game-changer for building high-performance REST APIs. In this video, we build a real-world Spotify Clone from scratch, covering: Modern Python Tooling: Setting up virtual environments and project structures [09:15]. FastAPI Essentials: Creating your first server, routing, and dynamic path parameters [16:16]. Data Validation: Leveraging Pydantic schemas for cleaner, safer code [02:57]. Database Integration: Connecting to PostgreSQL using SQL Alchemy ORM [02:44]. Security & Auth: Implementing JWT (JSON Web Tokens) and protecting routes [03:56:07]. Auto-Documentation: Using the built-in Swagger UI to test your endpoints instantly [04:58]. Watch the full course here: https://lnkd.in/dugZbs7u I’ve designed this to be a hands-on experience, starting from a "Hello World" to building complex relationships and authentication [03:57:56]. If you find the video helpful, please consider subscribing to the channel and leaving a comment—it really helps the algorithm reach more developers like you! Happy coding, Haider Malik

So I recently encountered another shocking and sad accident in my code, which is the race condition introduced by using singleton. I used to think singleton is good for storing things that only exist one copy in the system and can be reused. I saw a few of programs store their Redis connection pool in a singleton class, which looks nice. Then bad thing happens when I use that SDK I wrote in a Flask service with multithread enabled and save user states inside singleton. Then race condition happens, before one session finished, another already came in and overwrote the states. I would say from now I would really be extra meticulous about any designs and patterns I feel good about and then judge them before I put them in any running services. And for the solution? Actually it was quite straightforward, just save the states in each client object, no other tricks. Here is an example from Supervisely, which is an MLops platform which also support running services with their SDK as part of the code. https://lnkd.in/gc9bD8U7 #python #race_condition #singleton

New Blog Post 😃 In this #CopilotStudio post I share how to make your own MCP Server using Python 🐍 If you’ve never created an MCP Server before you might be surprised how straightforward the building process can be 🏗️ https://lnkd.in/gQav32d4

Clean code is good. Efficient queries are better. 🚀 One of the biggest lessons in Django backend development: performance is our responsibility, not the database’s problem. Understanding when to use: ✔️ "select_related()" ✔️ "prefetch_related()" ✔️ Query count monitoring ✔️ Proper DB indexes Small ORM decisions → huge production impact. Avoid N+1 queries early, and your future self will thank you. 💻⚡ #Django #Python #BackendDevelopment #PerformanceOptimization #DjangoORM #SoftwareEngineering

Built a portfolio that actually runs code. 💻✨ I got tired of static resumes, so I architected a fully interactive developer portfolio and paired with AI to build it out. Coolest Features: 🟢 Live Python Compiler: Write and execute code right in the browser (completely sandboxed!). 🟢 System Inspector: Watch the backend requests happen in real-time as you navigate the site. 🟢 Dynamic CMS: Built my own admin panel to manage projects, certificates, and messages. It's live and running on Render! Drop a "print('Hello World')" in my compiler and let me know what you think. Link : https://lnkd.in/gPX4-myZ #Python #Backend #DevPortfolio #SoftwareEngineering #Flask #odoo #sql #render #tech

I almost rewrote my entire Python backend in Rust this weekend. The problem: generating 300 API calls through Claude was taking 5 hours. One at a time, 60 seconds each. Sequential. Painful. I started researching multithreading. Connection pools. Async Python. Rate limit management. Retry queues. Then I found Anthropic's Batch API. You submit all your requests in one call — up to 10,000 at once. Anthropic handles the parallelism, the rate limits, the infrastructure. You get results back within an hour. And it costs 50% less than standard API pricing. My entire pipeline went from this: 1. Manage threads 2. Handle rate limits 3. Retry failures 4. Monitor progress 5. Pray nothing crashes To this: 1. Submit batch 2. Save the batch ID 3. Go make coffee or take a nap 4. Come back and download results If my script dies mid-wait? Doesn't matter. The batch lives on Anthropic's servers. Grab the ID, check back later. Results stay available for 29 days. Stack prompt caching on top and costs drop even further. No Rust. No async. No infrastructure. Just one API call I didn't know about. The lesson: before you optimize your code, read the docs. The answer might already be built.