ALL SPEAKERS
AT A GLANCE
Role
Senior Developer
Organisation
Team Internet
Category
Community
ABOUT THE SPEAKER
I’m Christian, backend developer and staff engineer with a focus on functional programming and data pipelines. I’ve worked on e-commerce platform and advertising systems handling terabytes of data with Elixir.
My interest in programming languages and the interaction of natural languages lead me to work with a wide range of database systems and programming languages. From interfacing C libraries as http endpoints in Rust to using AWS services to build data pipelines.
TALK
How to coach RAGs
Semantic search using sentence transformers and vector databases is very useful for giving agents the specialized domain knowledge they need. But overly broad results and large documents can stuff the context with too much irrelevant information. So how can you coach your semantic search to perform better?
In this talk, I present two ways of giving vector-search based systems a filter, a kind of persona that can be applied to any query to shift its results. After a brief introduction to sentence transformers, you’ll learn how to use vector operations to shift queries. With that knowledge, we’ll build a system that takes a small training set to construct an offset that can even work with untagged data. The second system will utilize a similar approach, but based on a hierarchical classification, to help reduce stray results and focus on topics actually relevant to the current context.
MORE VOICES
Other speakers at code.talks
See them live — and 100+ more.
One ticket, every track. Nov 4–5, 2026 at Kinopolis Hamburg.
Get your ticket





