Skip to content

NLP classifiers in high throughput applications - Real Networks

Photo of Ivan Borko
Hosted By
Ivan B.
NLP classifiers in high throughput applications - Real Networks

Details

Meetup topic

In this technical meetup, we will discuss a couple of particular challenges that modern NLP solutions have to handle when applied in extremely high throughput and low latency environment. We will talk about which technologies we used to get the performance benefit of Transformers against LSTM models while retaining the throughputs, and make Transformer based architectures a reality for our deployed A2P SMS content category classification models. We will also showcase some things we do with URLs within our data within and outside such environments, since URLs are a salient feature of most A2P messaging.

Presenters

Igor Mijić is the R&D Manager for the Zagreb R&D team for the KONTXT product by RealNetworks, a NLP AI-based product that provides real-time content monitoring and analysis capabilities to SMS channels. Shortly after finishing his Masters in Electrical Engineering, he continued his studies while working as a Research Assistant at the Faculty of Electrical Engineering and Computing, with a focus on speech processing, and applications of affective computing like fatigue & stress detection, cognitive load classification, etc. After moving to RealNetworks, he's switched focus to Natural Language Processing and practical implementations of research solutions. When he's not clicking and clacking at his keyboard, you can find him on a bike or in his kitchen.

Although Miljenko Šuflaj just recently completed his Masters in CS on FER, he's been an intern at RealNetworks for more than 3 years prior. Initially interested in Computer Vision he fell in love at first sight with Natural Language Processing, and currently works as a Junior Applied Research Scientist at RealNetworks. He is most passionate about building and optimizing the systems that process the data we analyze. In his free time he cooks, works out regularly and is currently designing and implementing a programming language.

Join us at the meetup!
When: 8.11.2022 (Tuesday), 18:00
Location: BIRD incubator, Ul. Franje Petračića 4, 10110, Zagreb
Duration: 45 min presentation + Q&A

Meetup will be held live. Reserve your spot through the meetup platform.
After the presentation, you are invited to hang out with us (drinks and food powered by Real Networks).

Photo of Data Science Croatia group
Data Science Croatia
See more events
BIRD Incubator
Ul. Franje Petračića 4 · Zagreb