Build MCP Servers with Python Starter Kit

Ended up packaging my F1 MCP setup into a small Python library which y'all can install :) 👉🏻pip install fastf1-mcp It runs as a local MCP server over FastF1, exposing race data as tools, so you can query sessions (2018+) directly from actual telemetry, timing, and strategy data. Works with Claude or any MCP-compatible client: just install, plug it into your config, turn the connector on and ask away! PyPI release: https://lnkd.in/gam9iSdk GitHub: https://lnkd.in/gkUZtfrf #F1 #Formula1 #MCP #Pypi

This week I shifted from maintaining RISC-V forks to pushing riscv64 support directly upstream. Five PRs merged in major projects: Docker CLI now includes riscv64 in its release build target. SQLAlchemy merged riscv64 wheels the same day I submitted the PR. pyzmq and jiter landed too. And wit-bindgen shipped riscv64 binaries in v0.54.0, completing the WASIP2 toolchain for RISC-V. Beyond merges, DuckDB now has a RISC-V page on their "Everywhere" site (https://lnkd.in/eaffG67n). I ran their full query engine on a BananaPi F3 board and it just works. Podman also builds from source in 18 minutes on the same board. On the infrastructure side, RISE (RISC-V Software Ecosystem) is providing 50 native riscv64 GitHub Actions runners for open source projects. numpy's lead maintainer is evaluating the runner app for their org. Apache Arrow maintainers are looking at it for the Apache GitHub organization. This solves the biggest blocker for riscv64 CI: maintainers telling me they won't ship binaries for architectures they can't test. I tested the full numpy test suite on native riscv64 hardware: 47,448 tests pass (99.97%). The 3 failures are not riscv64-specific. 30+ upstream PRs are open across the Python ecosystem. The goal: make pip install work the same on riscv64 as everywhere else. What open source projects would you most like to see with riscv64 support? #RISCV #RISCVEverywhere #OpenSource #Python #DataScience #DevOps #devEco #DockerCaptain

I typed a message in one terminal... and it appeared in another. Sounds simple, isn't it?  But understanding why it works made Networking finally make sense to me. 😊 Tried making a small TCP chat project using Python's socket library. Two terminals, two programs, messages flying between them in real time. No frameworks. No heavy libraries. Just raw sockets. The setup: 1. server.py : binds to port 9125, waits for a connection, receives and replies to messages. 2. client.py : connects to the server, sends messages, waits for replies. Here's what I learned : --> Every network connection needs an IP address and a port , just like your home address and house number. --> Networks only speak bytes , so you encode before sending and decode after receiving. --> The server blocks and waits. The client initiates. That's the client-server model. Small setup. Big fundamentals.  #Python #Networking #TCP #ContinuousLearning #LearningsomethingNew

🐍 Did you know Python Notebooks support %run? Most people treat notebooks as single-file scripts. But %run lets you break your code across multiple notebooks — and still run everything as a structured, organized pipeline. No cluster needed. A single node. Two cores. Full cost savings intact. 🔹 Modular code structure 🔹 Reusable notebooks as building blocks 🔹 Cost-efficient execution without sacrificing organization You get the best of both worlds: clean architecture AND lean compute. Stop cramming everything into one notebook. Structure your code. Keep your costs low.

484,000 engineers starred this repo. Most of them probably haven't opened it yet. It's called "Build Your Own X" and the concept is almost embarrassingly simple, Instead of just USING Git, Docker, Redis, or a web browser — you BUILD them from scratch. Not to ship But Just to understand. There's a Feynman quote at the top of the repo that says everything: "What I cannot create, I do not understand." That one line is the whole philosophy. The repo covers 30+ categories — build your own database, neural network, operating system, programming language, web browser, Git. Step-by-step. In Python, Rust, Go, C, JavaScript. You don't have to finish any of them. Getting halfway through building your own shell will teach you more than six months of YouTube tutorials. It's the difference between knowing how to drive and knowing how an engine works. One is a skill and the Other one is understanding. Repo Link: https://lnkd.in/gFtWc9zg #SoftwareEngineering #Programming #LearnToCode

This batch has a fun split: hard, measurable wins (memory from gigabytes to megabytes) sitting right next to the messy reality of letting agents and LLMs touch real systems. If you are juggling reproducibility, CI cache sanity, and tool safety, there is a lot here worth poking at. 🧩 A new chapter for the Nix language, courtesy of WebAssembly 🏗️ Cracking the Python Monorepo 🧰 Designing MCP tools 🤖 I built a programming language using Claude Code 🦀 Using Rust and Postgres for everything: patterns learned over the years 🗺️ Why value streams and capability maps are your new governance control plane VarBear is out. Read the issue: http://from.faun.to/r/rgXW Subscribe here: https://faun.dev/join/

Dependency management in Mage Pro just got a serious upgrade. ⚡ Mage Pro now supports uv as the primary package installer for virtual environment management. If you've used uv before, you already know what this means for installation speed. Faster installs. Isolated environments. Consistent execution across your pipelines. 🔒 No more environment drift. No more redundant installs. Just dependencies that work. What this means for users: ✅ No manual environment setup. ✅ No version conflicts slowing you down. ✅ Just fast, reproducible Python environments inside Mage Pro. Your pipelines just got faster. 🚀 🔗 Full docs here: https://lnkd.in/ecd_ttGi Less friction. More shipping. #dataengineering #python

🚀 MY DSA LEARNING LOG "Not a 100 days or 365 days challenge, just about my consistency in problem solving." Today I solved Trapping Rain Water – LeetCode 42 (Hard). ⚡ Approach 1: Prefix & Suffix Arrays ⏱️ Time: O(n) | 📦 Space: O(n) ⚡ Approach 2: Two Pointers (Optimized) ⏱️ Time: O(n) | 📦 Space: O(1) ⚡ Runtime: 2 ms ✨ Key Takeaway: Optimized space by moving from extra arrays to a two-pointer approach based on min(leftMax, rightMax). 🔗 GitHub: https://lnkd.in/g9Mgu4eT #DSA #LeetCode #ProblemSolving #Java #CodingJourney

🚀 Day 19 of Consistency | #75DaysLeetCodeChallenge 🧠 Today’s Problem : Implement Queue using Stacks (#232) 💡 Key Learning: This problem shows how to simulate one data structure using another, specifically implementing a queue using stacks. ⚡ Approach: Use two stacks → s1, s2 push → add element to s1 pop/peek → transfer elements from s1 to s2 if s2 is empty Then perform pop/peek from s2 empty → check both stacks 🧠 Why this works: Reversing stack order gives queue behavior (FIFO) Amortized O(1) time for operations Efficient use of stack properties 🔥 Result : ✔️ Runtime: 0 ms (Beats 100%) 📈 Great problem to understand how data structures can be built on top of each other. Consistency is compounding. Keep going. 💪 #Day19 #LeetCode #DSA #CodingJourney #100DaysOfCode #Python #Stack #Queue #Consistency

Most agent stacks depend on layers of tools. I wanted something simpler. So I built 𝗚𝗼 𝗔𝗴𝗲𝗻𝘁 𝗦𝗸𝗶𝗹𝗹 𝗥𝘂𝗻𝘁𝗶𝗺𝗲. 114 built-ins for HTTP, files, crypto, regex, JSON. Compiled in Go. Distributed as 𝗼𝗻𝗲 𝗯𝗶𝗻𝗮𝗿𝘆. No Python. No packaging. No runtime surprises. Build once → run anywhere. Your environment. Your control.