Build yourself flowers | ✰Vicki Boykis✰ Build yourself flowers<br>Apr 20 2026<br>This is an edited transcript of the keynote I gave at the Applied Machine Learning Conference in Charlottesville, VA in April 2026.I first wrote a draft of this talk by hand. This part took 2 months.<br>I then recorded myself giving a version of this talk with MacWhisper, and transcribed it with Whisper locally. This part took 45 minutes (the total time of my practice run.)<br>Then, I ran it through Gemini Flash 2.5 running in Pi to break into paragraphs. I also had Gemini break up my slide deck from a PDF I generated from Google Slides, into individual images to insert into the blog post and optimize the image format to webp for blog rendering. This part took about 10 minutes.<br>Then I went through the text paragraph by paragraph manually to make corrections, remove redundant phrasing and pauses, and added clarifications to make it more legible than a talk. This part took 3 hours.<br>Before all that, I generated the content for this talk. This part took 13 years.

I&rsquo;m Vicki, and I build machine learning systems.<br>I debated for a long time how to introduce myself. Am I a data scientist? Am I still a machine learning engineer? Am I an AI engineer now? I&rsquo;m not really sure. I think, like a lot of people over the past six months in the industry, I&rsquo;ve been having existential angst. So, I&rsquo;ll go with &ldquo;I build machine learning systems.&rdquo;<br>I&rsquo;ve built and broken systems at Tumblr, at Automattic, at Duo, at Mozilla.ai. Now, I build realtime personalization and search systems at Malachyte.<br>We&rsquo;ve had a lot of different conversations in all of the corners of the internet about how we should incorporate LLMs into machine learning workflows. And the question that came to me is, if all we&rsquo;re doing is generative AI, where does traditional machine learning even fit in here? Are we still doing machine learning at all? So, the larger question behind my existential angst is, what is the state of machine learning engineering as an industry today? That is - is it still worth doing machine learning engineering?<br>And the second question that came to me was, not only is it still worth doing ML, but, in an era where we&rsquo;re having LLMs generate a lot of code, when the most important thing is for us to ship quickly, to get to a prototype quickly, is it still worth doing machine learning well?<br>So hopefully we&rsquo;ll be able to definitively solve all of my existential crises in 45 minutes, but until then, I wanted to take a step back and talk about flowers.<br>In 2024, the lovely folks at PyData Amsterdam, invited me to give a keynote. I called it Build and Keep Your Context Window. ChatGPT was still fairly new, out for just about a year and a half at the time, and everyone was freaking out. I talked about how important it is to build and keep your own context window - the historical understanding of software tools and concepts, so that we can use the new tools and understand where they fit into the context of existing software engineering.<br>The analogy I used was the ChatGPT UI. At the time, everybody was using the text interface, where you have a sidebar and you have the big text box. All your chat sessions live in the sidebar, and the sessions have titles. And I talked about that as being part of a context window.<br>The context window is the amount of text that a model can recall at any given time, as measured in tokens. If the text you input overflows the context window, the model loses the thread of the conversation and can&rsquo;t reply coherently anymore. What determines the context window in reply is the attention mechanism in the Transformers model. And the attention mechanism is really a cache, which is really a hash map. So if you understand all these four things and how they fit together as concepts that naturally emerge over time successively, it&rsquo;s not perhaps as surprising that we get the emergence of LLMs.<br>Before I gave this talk, I got to spend a few days in Amsterdam (I was told they were the only two sunny days in the Netherlands in September) and particularly in the Rijksmuseum, which is the national Dutch museum of art.<br>It&rsquo;s absolutely enormous and beautiful. You could spend days there. But I only had a few hours, so of course I went to see the headline piece, which is the Night Watch by Rembrandt. When I went to see it, it was covered by this big plastic window and there was machinery set up all around it. I was thinking, what is this? Why is this blocking the painting?<br>When I went later to PyData Amsterdam, Robert Erdmann, who was working at the Rijksmuseum, gave an incredible talk about how he was using deep learning-based ink-removal, especially 3D imaging, pixel detection, and high resolution photography to see details in art at the museum that nobody was able to see before.<br>So for example, for The Night Watch, they set up all this equipment to be able to photograph it many times in many...