"The rise of AI has created something of a paradox for security professionals. On one hand, large language models and advances in machine learning mean that it's never been easier to parse through petabytes of data and craft agents that can spot and correct potential vulnerabilities. On the other, the ability for bad actors to exploit those ballooning piles of data has grown in tandem. So which wins out?"
"Founded by Elizabeth Nammour, a former security engineer at Airbnb, Teleskope takes a unique approach of using smaller large language models fine-tuned on specific problems such as detecting sensitive information within code files, rather than funneling everything into one larger model, which Nammour says allows her company to function more nimbly than competitors. "By segregating it into smaller problem sets, we are more accurate, but also faster," she told me."
Advances in large language models and machine learning both improve defenders' ability to analyze vast datasets and empower attackers to exploit sensitive information. Teleskope raised $25 million in a Series A led by M13, bringing total funding to $32.2 million, to help companies strengthen data security. The company was founded by Elizabeth Nammour, a former Airbnb security engineer, and builds specialized small LLMs fine-tuned for specific tasks like detecting sensitive data in code. The smaller, segregated models enable faster and more accurate detection compared with single large models. The product grew from proprietary tools created to locate, control, and delete sensitive customer data at scale. Competitors include AWS Macie, BigID, and Varonis.
Read at Fortune
Unable to calculate read time
Collection
[
|
...
]