Cheat code for MLEs to stand out in 2026

How to Break Into MLSys Through Open Source in 2026

TLDR: Open source is the cheat code to stand out in a saturated ML market. I’ll walk you through the best repositories to contribute to in LLM inference and RL infrastructure, plus some hard-earned tips on how to actually make meaningful contributions.

---

Introduction

Let’s address the elephant in the room: the ML job market is tough right now.

Positions are scarce, candidates are plenty, and everyone has “fine-tuned an LLM” on their resume.

So how do you stand out?

My answer has been the same for years: open source contributions.

Think about it. When you contribute to vLLM or PyTorch, you’re not just padding your CV. You’re building real skills, working with production-grade code, and — here’s the kicker — creating public proof of your abilities that any hiring manager can verify in 30 seconds.

I’ve seen folks get fast-tracked through hiring pipelines at top companies because a maintainer vouched for their contributions. That’s not luck, that’s strategy.

Today I want to give you a curated list of repositories worth your time if you’re interested in MLSys, specifically around LLM inference and RL infrastructure. These are active, impactful, and — importantly — welcoming to new contributors.

Let’s dive in!