Abnormal AI
@Abnormal

Abnormal stops cybercrime with AI. Our Behavioral Security Platform spans email, identity, and insider threat, powered by proprietary Behavioral AI.

 Las Vegas, NV
abnormal.ai
Joined July 2019
533 Photos and videos
Article

How we cut our coding-agent costs by 68% with prompt-based routing

Most teams route coding agents with a model router. We use a one-cent prompt instead, and it cut our cost per run by 68%. It runs once before each session and chooses the model, the reasoning effort,
3
7
277
Article

How we keep our knowledge base current

Our engineers ship, on average, about twice as many changes to production than they did a year ago. As AI accelerates velocity, the launch bottleneck moves downstream, and documentation usually breaks

1
8
14
760
Article

Friend or foe: teaching a model how to trust

How do you spot a threat that looks like a friend? Last month, an autonomous AI agent spent about two and a half days inside Hugging Face's infrastructure before anyone caught it. It wasn't a criminal
4
10
2,140
Article

Agent Reliability Is a Distributed Systems Problem

Building an AI agent stopped being impressive about a year ago. You can stand one up in an afternoon from just about anywhere. But when you put it in production where it’s making dozens of calls to

1
4
20
7,149
Article

Make Every PR Prove Itself

Authored by @mchatmajian Agents at Abnormal generate and request review for over 200 pull requests a day, and the green CI checkmarks tell us little about whether they are actually safe to merge.
7
16
3,512
Article

How We Made Multilingual Embeddings Work at 60,000 Queries a Second

Authored by @udayansarin and @meeraa7ramakrishnan Abnormal is in the business of stopping cybercrime—an industry where detection speed and coverage are the difference between stopping an attack and
6
22
4,763
Load more