Madalina Turlea will teach you how to build impactful AI features

I recently attended the "Build Al features hands-on" workshop hosted Catalina Turlea and Madalina Turlea at Netlight in Munich. Instead of just talking about Al potential, we dove straight into how product teams can move past simple vibe checks and build systematic evaluations using Lovelaice to validate feature accuracy and performance before deployment.

It was nice experience for me. I'm totally new to AI side. Even though I can easily follow the entire session without any difficulties. The main reason is it was completely "Interactive session". Both technical and non-technical persons participated in this session. Catalina was really humble to answer all our questions and clear our doubts during entire session. Rather than a typical talk along with presentation slides she showed us a real working scenario by using her product "Lovelaice". That was a mind-blowing, because everyone in that session got a chance to experience various results in real time. Is "Latest Model is the best one always" ? --- Catalina clearly answered this question with results of Lovelaice. I really appreciate Catalina and her team efforts behind building such nice product in AI industry. Expecting more workshops like this from your side.

I came in from the engineering side, kind of braced for a high-level "what are evals" overview — and it just wasn't that. By the end I had an actual eval suite on a real use case instead of a deck. What made it click was sitting with real model outputs and tagging exactly what went wrong — hallucinations, missing info, wrong format — and then building the checks around those. I'd been vibe-checking everything before and didn't realise how much I was missing. Paced well, and it all ran really smoothly.

It was super interesting to find out more about evals and how to get out of the manual testing. I will for sure take a lot from the course into my day to day. Madalina and Catalina are amazing teachers who never seem to be short of an answer. I highly recommend this course to anybody!!

It was especially useful for understanding the role of structured evals in building reliable AI products.