Python Weekly (Issue 748 June 4 2026)

Author

Rahul Chaudhary
June 04, 2026

In partnership with

200+ Claude Prompts Top Professionals Actually Use at Work

Claude can be your analyst, editor, and strategist.
But most professionals are using it to fix grammar.

These 200+ Claude prompts take it from grammar tool to your most powerful AI work assistant.

Sign up for Superhuman AI and get:

  • 200+ ready-to-use Claude prompts to get real work done in minutes — researched, tested, and used by professionals at Google, Microsoft, and NASA

  • Superhuman AI newsletter (4 min daily) so you keep learning new AI tools and skills to stay ahead in your career — the prompts are just the beginning

Welcome to issue 748 of Python Weekly. Let's get straight to the links this week.


Articles, Tutorials and Talks

A tour through AI engineering for developers who already know how to ship software. Fourteen chapters, no LinkedIn voice, no slow warm-up. We will go from 'what is a foundation model' to 'how do you run agents in production on Google Cloud' without skipping the parts that matter.

The article introduces django-integrity-policy, a package that helps Django applications enforce browser Integrity Policy headers, protecting against unauthorized or tampered third-party scripts and resources.

Learn to build robust web scrapers that can defeat modern anti-bot systems. In this full-stack course, you will transition from basic Python scripting to deploying a full MERN dashboard that scrapes and visualizes real-world data from Amazon, Booking.com, Indeed, and the TIOBE Index.

The article explores practical techniques for handling circular imports in Python while preserving static type safety, using a financial modeling framework as a real-world example. It argues that strategically placed local imports and dependency thunks are often the least-bad solution, offering better type checking and maintainability than context dictionaries, builder patterns, or large dependency manifests.

The video demonstrates how to add webhooks to a FastAPI application and then evolve the design toward an event-driven architecture that decouples integrations from API endpoints. By using events to handle side effects, the approach makes applications more modular, extensible, and easier to maintain as new integrations and workflows are added.

Using PyTorch custom operations is common in PyTorch models. PyTorch custom operations can be custom classes and custom functions implemented in C++ and CUDA and used in both Python and C++ inference programs. This post shares how to implement PyTorch custom operations in C++ and CUDA, and how to use them in PyTorch models and AOTInductor compiled inference programs, using a simple identity convolution example.

Learn how servers actually work by building a tiny TCP server and client from scratch. A hands-on introduction to sockets, TCP, and the network programming model every backend, DevOps, and platform engineer should go through at least once.

In this video, we'll look at new features coming in the Django 6.1 release, including fetch modes, mailers settings, and new database-level on-delete functionality for foreign key relationships.

This article introduces Pattern Screamer, a reconnaissance tool that uses multi-protocol TTL tracing to discover routers, subnets, and network topology in environments with unknown addressing schemes. By combining heuristic subnet sampling with ICMP, TCP, and UDP probes, it dramatically reduces scan volume while enabling attackers and administrators to map network infrastructure without exhaustive address-space scanning.

Plugin Playground lets developers prototype, test, and iterate on JupyterLab extensions entirely from within JupyterLab, eliminating much of the setup and packaging overhead of traditional extension development. The article follows the journey from a simple idea to a publishable extension, showing how Plugin Playground accelerates experimentation and lowers the barrier to building custom JupyterLab functionality.

Agentic RL: Token-In, Token-Out Done Right

Training agentic LLMs with reinforcement learning can silently break when tool-calling workflows re-tokenize model outputs between turns, causing gradients to be applied to sequences the model never actually generated. The article introduces the Token-In, Token-Out (TITO) approach, which avoids re-encoding decoded tokens and preserves training correctness while simplifying multi-turn RL for tool-using agents.


Interesting Projects, Tools, and Libraries

AI agent to evaluate and score resumes.

The best way to use Hermes Agent from the web or from your phone!

A platform for reproducible world model research and evaluation.

High-fidelity world models for general embodied intelligence, such as data engines and world simulators.

Build, evaluate, and integrate long-term memory for self-evolving agents.

A codebase intelligence toolkit that finds dead code, analyzes change impact, scores complexity, detects duplication, and generates interactive codebase graphs.

Small FastAPI backend that uses a Rust pyo3 extension for AI token-aware rate limiting backed by Redis.

A lightweight Python CLI + API for managing servers over SSH. Add hosts, run commands, transfer files, target groups by tags - no Ansible DSL required.

alembic-git-revisions: Stop fighting Alembic's "Multiple head revisions are present" error. This library drops the hardcoded down_revision and chains migrations from git commit history instead, so parallel branches never collide and the chain stays linear regardless of merge order. Apache-2.0, works alongside existing migrations.


New Releases

Django 6.0.6 and 5.2.15 patch five low-severity security issues, including signed-cookie validation flaws, potential unencrypted email delivery after failed STARTTLS handshakes, and several caching bugs that could expose private data


Upcoming Events and Webinars

There will be following talks

  • Kicking Your Robot Out Of The Nest

  • Luna, Mochi, and the Graph of Seattle

  • Weather or Not : Full Stack Data Science from Cornfield to API

There will be a talk, Mini Shai Hulud is no longer just an npm problem.

There will be a talk, Efficient Neural Networks Through Tensor Networks.

There will be following talks

  • The AI Adoption Paradox: When Technology Moves at Internet Speed and Organizations Move at Paper Speed

  • Am I still a Good Coder?

There will be following talks

  • A Match Made in Heaven: LLM Judgment at Vector-Search Speed

  • Dr. Data: How We Learned to Stop Worrying and Love the AI Bomb

  • Cost-Efficient AI Systems in Practice


Our Other Newsletters

Programmer Weekly - A free weekly newsletter for programmers.

Founder Weekly - A free weekly newsletter for entrepreneurs featuring best curated content, must read articles, how to guides, tips and tricks, resources, events and more.