I have about 30 years of programming experience, mainly with C# and C++, and I've experimented with Python a little. With the growing importance of AI, I'd like to become comfortable with Python so I can work with tools such as LangGraph.
I already understand many AI concepts because I've built custom solutions and used tools such as Stateless and Temporal, but I've never started a Python project from scratch with a solid development workflow, including project structure and CI/CD. I've also been using Cursor, which raises the question of how much Python I need to know personally versus how much I can delegate to AI.
I'm not comfortable blindly trusting generated code, though. I want to understand what is happening at each level and learn Python in a way that follows Python-specific conventions rather than simply recreating C# or C++ patterns. What learning resources, tools, workflows, and practical advice would you recommend for an experienced developer making this transition?
4 Answers
A good book or structured course can provide a much better foundation than passively accepting code from an LLM. Use AI as a supplement: ask it to clarify a confusing explanation, provide another example, or review an approach after you’ve tried it yourself. The stronger your fundamentals are, the better you’ll be at checking generated code, controlling the architecture, and avoiding unnecessary technical debt.
For a practical project setup, tools worth investigating include uv for Python versions, virtual environments, dependencies, and lockfiles; Ruff for formatting and linting; Pyright for static type checking; pytest for tests; and Rope for refactoring. These tools provide useful guardrails for both humans and AI-generated code. Python’s interactive REPL is also valuable for quickly exploring language features, libraries, and standard-library behavior without turning every experiment into a full program.
Try not to judge Python only by the amount of behavior it hides. Higher-level abstractions can be a good tradeoff when the problem does not require low-level control, and many Python libraries hand performance-critical work off to native code anyway. Learn enough about the runtime and ecosystem to understand the tradeoffs, but also embrace the situations where Python’s flexibility and short feedback loop are useful.
The basic syntax should be fairly easy for an experienced programmer, especially since Python is higher-level and has less ceremony than C# or C++. Mastering the ecosystem, idioms, packaging, testing, and deployment practices will take longer, just as it would with any language. A Python-focused IDE can make the transition smoother, and AI integrations are optional rather than required.
Start with the official Python tutorial rather than an introductory course aimed at complete beginners. It should let you move quickly through familiar programming concepts while highlighting Python’s own data model, syntax, and standard library. Then build a small project without having an AI write everything for you so you develop an instinct for how Python code is normally structured.
That makes sense. I’m particularly interested in learning the idiomatic approach instead of translating my usual C# patterns directly into Python.
The main benefit of those tools is consistency. They make it easier to keep a project predictable while you learn, rather than relying on convention alone.