The Agony of Evaluating Agents Wrongly

You know that gut-wrenching feeling when you deploy a seemingly perfect agent system, only for it to crash and burn in a live scenario? I’ve been there too many times. It’s like investing in a hamster to defend your fortress. Useless. I remember back in October 2022, I deployed

Dapo: An Open-Source LLM Reinforcement Learning System at Scale

As an ML engineer, I’ve seen firsthand the challenges of fine-tuning large language models (LLMs) for specific tasks. While supervised fine-tuning (SFT) is effective, it often falls short in aligning models with complex human preferences or nuanced real-world reward signals. This is where reinforcement learning from

Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference

By Alex Petrov, ML Engineer

Seed Diffusion marks a significant step forward in generative AI. It’s a large-scale diffusion language model built for practical applications, prioritizing not just the quality of output but also the speed at which it generates that output. This article explores the

Performance Engineer – Deep Learning: Practical Strategies for ML Optimization

As an ML engineer, I’ve seen firsthand how critical performance is in deep learning. Models that are brilliant in theory can fail in practice if they’re too slow, too resource-intensive, or prone to instability. This is where the “performance engineer – deep learning” role becomes

Model Optimization: Real Talk on Fixing Bad Habits
Ever spent weeks training a model only to find out it runs slower than your grandma with a dial-up internet connection? Let me tell you, I’ve been there, and it’s a pretty frustrating place to be. But here’s the kicker: most of these issues aren’t about fancy

Let me tell you about the time I nearly threw my laptop out of the window. It was 2025, 3 AM, and I was stuck trying to optimize an agent system that just wouldn’t cooperate. Seriously, it felt like a stubborn mule refusing to move an inch despite all the coaxing, poking, and

Model Optimization: Stop Rolling Your Eyes and Do It Right

Let’s talk about model optimization, and yes, I know. You’re rolling your eyes because it sounds boring, tedious, or maybe you’re thinking, “I don’t need this; my model is already doing fine.” Well, hang tight. Years of building agent systems have seasoned me with frustration (and

Model Optimization: Real Talk for Better Performance

Alright, folks. Let me kick this off with a little story. A couple of years ago, I got stuck optimizing a recommendation system that had more parameters than a high-end gaming rig. It was a mess. The problem? Everyone was fixated on stacking layers and increasing the complexity when