Notes from real systems.
Practical write-ups on backend engineering, AI systems, and the decisions that hold up in production — not generic tutorials.
Why your cache layer fails under load
Most teams add Redis and call it done — until stampeding herds and hot keys show up.
The fix is rarely more nodes. It is request coalescing, TTLs with jitter, and ownership.
- 9 articles
- Published on this page
- Topics
- Backend, AI & architecture
- Filter below
- Browse by category in the list
Everything published so far.
A RAG chatbot, line by line in Python
Walk a small RAG chatbot in Python in the order it runs: four methods once, four on every question. Then three live runs, and the three gaps they leave.
RAG in 8 lines, and where memory goes
The eight stages of a basic RAG chatbot, one line of code each. Why the closest chunks are not the answer, and the one line that changes to give it memory.
A sure answer is not a checked one
Why ChatGPT can sound sure and still be wrong, what RAG's three steps do, why you still open the source, and when a changed page needs RAG instead of fine-tuning.
Software job posts, five years
US software job posts doubled by 2022, crashed, and are still 22% below early 2020. AI did not replace developers, but the rebound is mostly senior and AI-titled roles.
The Origin header
Origin names the site that sent the request. CORS can hide the answer from that site’s JavaScript. CSRF is another site using your Cookie.
The Cookie header
The browser adds Cookie on the page. You do not set it in fetch. Postman can still send it if you paste it. It is not Authorization.
Token in the header
The blog JSON is the post. The bearer token proves who is asking, so it stays in Authorization, not in the body.
Request headers
Name Content-Type, Accept, and Authorization, use each on a blog API, then catch the token-in-the-URL mistake.
Five request verbs
Name GET, POST, PUT, DELETE, and PATCH, use each on a blog API, then catch the /delete path mistake.
Materials you can learn from on your own.
Articles, video tips, and project write-ups — free content to read, watch, and revisit whenever you need it.