Martín Quesada

Tag: English

Cascading pipelines with DSPy are kind of good

Posted:   |   Categories: article   |   Tags: English

The global machine learning race has given us many great things: stupidly high GPU and RAM prices, a great pretext to fire workers while companies pretend that we are not in a global recession (AI is making us so productive! Also please do NOT look at my last quarter), and some truly good models. Although we are most likely not getting AGI, as seemingly every major AI company is converging on creating the best video slop website that money can buy, which is totally what would be reasonable to do when you are close to obtaining computer god. Read more...

An unhelpful summary of PyConES 2024

Posted:   |   Categories: article   |   Tags: English

Last week I went to Vigo to attend the Spanish PyCon, PyConES. I traveled there fully sponsored by Datamaran, the company that I work for. I attended as a speaker, since a colleague and I prepared a talk proposal that got accepted (something something ESG with clustering and LLM reformulation). Day 1 🌦 The first day had only two turns for workshop tracks. I attended: Overcoming the One Billion Row Challenge with Python (Jordi Contestí, Kiko Correoso, Ernesto Coloma Rotger) This workshop showed different tools and approaches that can be used in Python to overcome the somewhat known One Billion Rows Challenge. Read more...

Taking a look at (some) tokenizers

Posted:   |   Categories: article   |   Tags: English

Recently I have been working on writing a tokenizer from scratch in Rust. In the process, I wanted to really understand the implementation of some commonly used tokenizers. Fun, I know. Moses Tokenizer Webpage http://www2.statmt.org/ and the implementation that I’ll talk about tokenizer.perl. If you already know Moses, you know. Silent nod. For everyone else, Moses is an NLP framework written in Perl, focused on statistical machine translation. It is super easy to install anywhere and everyone loves it because of it. Read more...

An unhelpful summary of AMTA 2022

Posted:   |   Categories: article   |   Tags: English

Last September I attended virtually the 15th biennial conference of the Association for Machine Translation in the Americas, A.K.A. AMTA 2022. There I presented an adaptation of my master’s thesis in collaboration with my tutor (shameless self-promotion here). Even though at the moment I am not enrolled in any course, he helped me with several revisions and provided the budget for attendance, so I wanted to thank him before I move on. Read more...

Things I have read (I)

Posted:   |   Categories: journal   |   Tags: books, English

This is the first of a series of posts where I’ll write some impressions on the latest books that I have read. It started as a personal log, but since I tend to read a ton of book reviews and they are my main way of discovering new stuff, I reckon that someone who reads it may also get something out of it. The Grownup I’m currently in Seoul for a while, and my choices for new stuff to read are limited to second-hand bookstores that have an English section, so not much reading in Spanish will be done until I am back home. Read more...