Hi, HN community. I'm Ines and I work at Sword Health, a company operating in the AI for Healthcare space. We have served close to 1M people, helping them recover from physical pain, pelvic issues, mental health issues, and others, from the comfort of their homes.
We have recently built Blade Code - the star of this article - within our engineering team to face 3 things:
> AI usage was growing faster than we could control it
> costs were also sky-rocketting
> yet we couldn't associate either growth with concrete outcomes (specific features built, direct impact on the business, etc.)
Blade Code is an LLM control plane that allows us to manage routing, access, budgets and cost attribution all into one layer. It allows us to route AI usage within our engineering orgs across teams and providers. And it helped us decrease our costs with LLMs in 94% MoM (we go over the cost methodology in the article).
Our CTO, Jorge, is really the mastermind behind this journey, and I've been asking him to join the HN community and share the story, as I think it can be helpful for other engineering organizations. He agreed on sharing the story - the article of the post -, but as for sharing it with HN folks he let that job for me as I am more of a HN fan (haha).
Would love to answer any questions you might have.
Hi, HN community. I'm Ines and I work at Sword Health, a company operating in the AI for Healthcare space. We have served close to 1M people, helping them recover from physical pain, pelvic issues, mental health issues, and others, from the comfort of their homes.
We have recently built Blade Code - the star of this article - within our engineering team to face 3 things: > AI usage was growing faster than we could control it > costs were also sky-rocketting > yet we couldn't associate either growth with concrete outcomes (specific features built, direct impact on the business, etc.)
Blade Code is an LLM control plane that allows us to manage routing, access, budgets and cost attribution all into one layer. It allows us to route AI usage within our engineering orgs across teams and providers. And it helped us decrease our costs with LLMs in 94% MoM (we go over the cost methodology in the article).
Our CTO, Jorge, is really the mastermind behind this journey, and I've been asking him to join the HN community and share the story, as I think it can be helpful for other engineering organizations. He agreed on sharing the story - the article of the post -, but as for sharing it with HN folks he let that job for me as I am more of a HN fan (haha).
Would love to answer any questions you might have.