2026
Life in the Slow Lane: Cost-Efficient Streaming Through Latency Tiering
Everyone loves real-time data until the cloud bill arrives. How splitting one fast pipeline into lanes made streaming cheaper.
40 min
On stage.
My two cents about topics that fascinate me, delivered live to real people around the world. Usually educational, sometimes entertaining.
2026
AI vs. Karen: Building Real-Time Sentiment Analysis Pipelines at Scale
How to spot the impossible customers before it's too late: real-time sentiment analysis with Kafka and Flink, and when an LLM or a small language model fits better.
50 min
Also given at AI Summit Europe 2025
2025
This Place is a Mess, Rewrite Everything!
You joined a new team, the legacy code makes no sense, and the urge to rewrite it all keeps getting louder. Should you?
15 min
Also given at Reversim Summit 2022 (lightning talk, in Hebrew)
On the page.
My thoughts on the things that keep my mind running, in long form. Occasionally a rant.
Words Matter: How Prompting Shapes Accuracy, Speed, and Cost
Even at the current generation of LLMs, the prompt still plays an extremely important role in performance for tasks.
On Technical Debt and Why We Can't Have Nice Things
Technical debt is usually created in return for gaining a certain value, even if you are not realizing it.
You Are The Problem: Stop Blaming Jira
Jira tickets don't kill productivity: People kill productivity.
In person.
If you made it this far, I might as well introduce myself.
I’m Josef, I also go by Yossi. I’ve been working
professionally as a software engineer and engineering leader for the past 18 years.
I got into engineering because I love how things work. I stayed because I love how people work. Turns out those two worlds have more in common than you’d think.
If you want to know more →