4 ms·

I use Foundry for work. It makes data ingestion, cleaning, quality check and automation easy. After all the data is ingested, running analysis/RAG on them becom

I use Foundry for work. It makes data ingestion, cleaning, quality check and automation easy. After all the data is ingested, running analysis/RAG on them become extremely easy. Basically, it's end-to-end data engineering and analytics. And the more a company uses/invests into the platform, the more benefit and locked-in they are.

hermitcrab 2y ago
RAG?
- mandevil 2y ago
  Retrieval Augmented Generation. https://en.wikipedia.org/wiki/Retrieval-augmented_generation https://en.wikipedia.org/wiki/Retrieval-augmented_generation Basically, using your actual data/documents to supplement a general purpose LLM and generate better answers for your specific use case.
alexpetralia 2y ago
"End-to-end data engineering and analytics" is quite a bold claim from a single service provider. Here is the link for anyone interested: https://www.palantir.com/platforms/foundry/ https://www.palantir.com/platforms/foundry/ and a YouTube explainer: https://www.youtube.com/watch?v=ZGGRCTTjLfQ https://www.youtube.com/watch?v=ZGGRCTTjLfQ Given you've used it, just how self-service is it? To me this seems like such a large claim that - if it's doable - I'm surprised there are not more competitors in the "vertically integrated data providers" space.
- maeil 2y ago
  > Given you've used it, just how self-service is it? To me this seems like such a large claim that - if it's doable - I'm surprised there are not more competitors in the "vertically integrated data providers" space. It is both very self service and not very self service. That's why they employ the FDE model from the article, to actually ingrain it into the client company to the point that it becomes self service. It's extremely hard to build such a product from scratch and have it actually be good, that's why there's no competitors. Especially providing the finely grained security controls that the article talks about, and have the platform be secure. There's a reason their security team wins the biggest CTFs half the time.
- robertkoss 2y ago
  It is completely self service by now. I have my own stack for testing purposes. Of course if you want to deploy this to an enterprise things will differ, but that is the same for Snowflake, Databricks etc.
justmarc 2y ago
It certainly sounds like they've created an excellent product both for its value to the customer as well as value to their shareholders. That's what companies should all be built and optimized to do. That's what it's about.

hermitcrab 2y ago

RAG?

mandevil 2y ago
Retrieval Augmented Generation. https://en.wikipedia.org/wiki/Retrieval-augmented_generation https://en.wikipedia.org/wiki/Retrieval-augmented_generation Basically, using your actual data/documents to supplement a general purpose LLM and generate better answers for your specific use case.

alexpetralia 2y ago

"End-to-end data engineering and analytics" is quite a bold claim from a single service provider. Here is the link for anyone interested: https://www.palantir.com/platforms/foundry/ https://www.palantir.com/platforms/foundry/ and a YouTube explainer: https://www.youtube.com/watch?v=ZGGRCTTjLfQ https://www.youtube.com/watch?v=ZGGRCTTjLfQ Given you've used it, just how self-service is it? To me this seems like such a large claim that - if it's doable - I'm surprised there are not more competitors in the "vertically integrated data providers" space.

maeil 2y ago
> Given you've used it, just how self-service is it? To me this seems like such a large claim that - if it's doable - I'm surprised there are not more competitors in the "vertically integrated data providers" space. It is both very self service and not very self service. That's why they employ the FDE model from the article, to actually ingrain it into the client company to the point that it becomes self service. It's extremely hard to build such a product from scratch and have it actually be good, that's why there's no competitors. Especially providing the finely grained security controls that the article talks about, and have the platform be secure. There's a reason their security team wins the biggest CTFs half the time.
robertkoss 2y ago
It is completely self service by now. I have my own stack for testing purposes. Of course if you want to deploy this to an enterprise things will differ, but that is the same for Snowflake, Databricks etc.

justmarc 2y ago

It certainly sounds like they've created an excellent product both for its value to the customer as well as value to their shareholders. That's what companies should all be built and optimized to do. That's what it's about.