Hear how contributing to open source and sharing your work publicly can lead directly to job opportunities in developer advocacy.
Discover how publicly analyzing and improving upon existing technologies can make you a highly visible and attractive candidate for top companies.
Understand why simply using larger context windows in models like GPT-5 is not a scalable or cost-effective solution for production applications.
A brief explanation of how text is converted into numerical vectors to represent its semantic meaning, enabling similarity searches.
Learn how to use a vector database with reference examples to classify text, avoiding costly LLM calls for simple categorization tasks.
Discover how to use semantic routing to direct user prompts to the correct function or to block inappropriate topics without involving an LLM.
Implement semantic caching to store and retrieve answers for semantically similar user questions, drastically reducing redundant LLM calls and improving response time.
Explore techniques like self-improvement, hybrid fallbacks, and prompt chunking to fine-tune and improve the accuracy of your semantic patterns.
Learn how to avoid common caching pitfalls, such as misinterpreting negation, by using specialized embedding models trained for semantic caching.
The discussion covers practical considerations like preventing stale cache data with TTL, managing data ownership, and how Redis handles persistence.
34:39
1:00:56Mary Grygleski
42:26Meta Atamel & Guillaume Laforge
58:00Vijay Krishan Gupta & Gauravdeep Singh Lotey
32:26Krzysztof Cieślak
57:52Rainer Stropek
59:38Erik Bamberg
28:57Dieter Flick & Michel de Ru
Jobs that call for the skills explored in this talk.
webLyzard
Vienna, Austria