HeyJaredData Science Espresso by Sarah Lea
Machine Learning, Deep Learning and Data Science simplified and served to your coffee break☕!
- Indexed issues, last 90 days
- 3
- Latest publication
- Sep 19, 2026
- Audience
- Checking…
- Earliest in this view
- Jul 13, 2026
Latest issues
From Word to Obsidian and Claude Code: How I Write Academic Papers Now (opens the original)
Read excerpt
I opened an empty folder in VS Code, started Claude Code and entered /obsitex:obsitex-init. After eleven questions, the folder contained a complete project for a seminar paper: section files with placeholders, a bibliography and a file with all the settings for the conversion. In the Obsitex web app, I selected the folder and then opened the PDF:<a class="image-link image2 is-viewable-img" href="https://substackcdn.com/image/fetch/$s_!0LNL!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fs
RAG or a 1M Token Context Window? (opens the original)
Read excerpt
When Kimi K3 came out in July with a context window of one million tokens, I asked myself the question a lot of people probably did: can I skip RAG now?A selection of what I have published over the last two years adds up to 127,068 tokens. This is twelve percent of the window. With that in mind, I ran a small experiment: I asked the same twelve questions, but once through a RAG pipeline with the top five chunks and once with all 32 articles in the prompt. The rest remained the same: I used the s
Building a Multi-Agent Support Ticket Triage With LangGraph (opens the original)
Read excerpt
Imagine it is Monday morning and 80 new support tickets are sitting in the inbox.Most of them are routine:a password reseta “how do I change my email” questiona billing question that comes up ten times a weekAnd somewhere in that pile sits the one ticket that actually matters:“I was charged twice and I need a refund today.”During my time as an IT consultant at an SME, this was one of the most tedious problems I constantly faced.This week, I rebuilt that exact workflow as a small team of AI agent
Publishing over time
Last 90 days. Choose a month to open its work.
Recurring subjects
Named in the text we hold. One piece can cover several.
Audience
No verified audience measurement yet.
About this data
Counts cover the work we have indexed. Tone needs enough text and a confident classification. Excerpts and episode notes are not full articles or transcripts.
Identity or attribution wrong? Suggest a correction.