flâneur

What's new with ML in production | ★❤✰ Vicki Boykis ★❤✰

vickiboykis.com · 4,917 words · saved by 1 readers

In 2023, I wrote two pieces on machine learning engineering for The Pragmatic Programmer. (Part 1 and Part 2). However, since I started working with LLMs recently, neural architectures have changed some of those assumptions. To be clear, most of machine learning in production is still not related to large language models or generative AI, and even deep learning projects, of which LLMs are a small subset, make up no more than 10% of the market, at most. But it’s still helpful to compare and contrast classical ML systems with their LLM counterparts, because this is where we’re going in the future. At its core, a machine learning model is an algorithm that performs compression on a corpus of data. In the first half of the 20th century, when digital communication in the then-nascent telephone industry exploded,scientists and engineers at Bell Labs started studying the projection and representation of information. In developing this field, scientists like Harry Nyquist,Ralph Hartley, and Cl

What's new with ML in production | ✰Vicki Boykis✰ What's new with ML in production Jan 15 2024 Image with some help from Dingboard. In 2023, I wrote two pieces on machine learning engineering for The Pragmatic Programmer . ( Part 1 and Part 2 ). However, since I started working with LLMs recently , neural architectures have changed some of those assumptions. To be clear, most of machine learning in production is still not related to large language models or generative AI, and even deep learning projects, of which LLMs are a small subset, make up no more than 10% of the market, at most. But it'

related reading