@kodekloud: Jev vs. ML Classifiers: What Actually Changed? Traditional machine learning models like Random Forests have sorted support tickets for years. However, adding a new category forces you to gather labeled examples and retrain the model from scratch. Modern structured AI models solve this by accepting plain English categories directly inside your prompt. When your needs change, you simply edit the prompt text rather than retraining the system. Because these models understand natural language context, they can even detect subtle cues like customer frustration without matching exact keywords. Unlike rigid classifiers that lock categories in during training, structured decision models evaluate custom options on the fly. This gives developers the flexibility of generative AI paired with the reliability of structured outputs. #MachineLearning #ArtificialIntelligence #SoftwareEngineering #TechExplained #AIEngineering #SystemDesign

KodeKloud

Open In TikTok:

Region: SG

Wednesday 23 September 2026 03:58:53 GMT

21124

1043

Music

Download

No Watermark .mp4 (0.7MB) No Watermark(HD) .mp4 (0.89MB) Watermark .mp4 (1.94MB) Music .mp3

Comments

luisdecker388 :

Job posts with 5+years of experience using jev coming soon

2026-09-23 17:56:02

Matthew Tam :

is Jev not just a zero shot classifier ?

2026-09-23 05:26:42

Lucian :

is laya better than jev?

2026-09-23 15:39:39

I’m your father :

So this just combines classifier with small llm. If you write text as input it must be processed somehow and this is rather something like small bert or something there and finally classification task instead of text generation.

2026-09-23 06:38:51

Derek :

Hype and nothing special. Classifiers have been around for years.

2026-09-23 15:08:54

Cannot say :

bruv heavy breathing is killing me 😂

2026-09-23 07:53:37

Michael Ramos :

theyre full 🙃

2026-09-23 14:50:52

senormat :

Awesome what did you use to animate this video

2026-09-23 04:18:52

śmæŕŧ :

only issue: with this little output it is impossible to assess the actual quality of the model; there is no reasoning - which makes the quality of the model even more important; and you have no idea if the model understood the task or filtered out the relevant information.

2026-09-23 09:36:35

UMNZ 🇳🇿 :

How does it work?

2026-09-23 04:43:58

Jimit Talekar :

someone copied my idea I had built this system for ERP

2026-09-23 05:33:12

AI & Coding eBooks :

I've made an ebook on Jev

2026-09-23 16:15:47

demonslayer :

i dont understand the hype, there no use for normal users

2026-09-23 05:40:39

TMM :

how was it trained?

2026-09-23 05:33:19

To see more videos from user @kodekloud, please go to the Tikwm homepage.