Basejump AI
@basejumpai.bsky.social
Basejump AI is an AI Data Analytics Platform that provides an AI assistant so anyone can chat with their data. No more waiting for weeks for data insights, using Basejump you can get answers in seconds and share them with your team!
3/3
When adding these to our Agentic RAG pipeline, the accuracy gets high enough to be used in real world applications.
What have you seen improves your text-to-SQL RAG accuracy?
When adding these to our Agentic RAG pipeline, the accuracy gets high enough to be used in real world applications.
What have you seen improves your text-to-SQL RAG accuracy?
March 3, 2025 at 6:46 PM
3/3
When adding these to our Agentic RAG pipeline, the accuracy gets high enough to be used in real world applications.
What have you seen improves your text-to-SQL RAG accuracy?
When adding these to our Agentic RAG pipeline, the accuracy gets high enough to be used in real world applications.
What have you seen improves your text-to-SQL RAG accuracy?
2/3
One of the ways to increase this accuracy is with checks on the SQL query itself. At Basejump, we use a process similar to the following to ensure users are getting correct results:
- Validate the primary keys in the result
- Validate the query joins
- Validate the query syntax
One of the ways to increase this accuracy is with checks on the SQL query itself. At Basejump, we use a process similar to the following to ensure users are getting correct results:
- Validate the primary keys in the result
- Validate the query joins
- Validate the query syntax
March 3, 2025 at 6:46 PM
2/3
One of the ways to increase this accuracy is with checks on the SQL query itself. At Basejump, we use a process similar to the following to ensure users are getting correct results:
- Validate the primary keys in the result
- Validate the query joins
- Validate the query syntax
One of the ways to increase this accuracy is with checks on the SQL query itself. At Basejump, we use a process similar to the following to ensure users are getting correct results:
- Validate the primary keys in the result
- Validate the query joins
- Validate the query syntax
3/3
The applications that drive the most value in knowledge work are those where the results are able to be audited. Full stop. 👀
www.sequoiacap.com/article/gene...
The applications that drive the most value in knowledge work are those where the results are able to be audited. Full stop. 👀
www.sequoiacap.com/article/gene...
Generative AI: A Creative New World
www.sequoiacap.com
February 24, 2025 at 6:37 PM
3/3
The applications that drive the most value in knowledge work are those where the results are able to be audited. Full stop. 👀
www.sequoiacap.com/article/gene...
The applications that drive the most value in knowledge work are those where the results are able to be audited. Full stop. 👀
www.sequoiacap.com/article/gene...
2/3
- Sonya Huang, Pat Grady, Generative AI: A Creative New World, Sequoia Capital
This is a bold statement from Sequoia is something to think about to say the least. It's an exciting time to be in the Generative AI space!
- Sonya Huang, Pat Grady, Generative AI: A Creative New World, Sequoia Capital
This is a bold statement from Sequoia is something to think about to say the least. It's an exciting time to be in the Generative AI space!
Generative AI: A Creative New World
www.sequoiacap.com
February 24, 2025 at 6:36 PM
2/3
- Sonya Huang, Pat Grady, Generative AI: A Creative New World, Sequoia Capital
This is a bold statement from Sequoia is something to think about to say the least. It's an exciting time to be in the Generative AI space!
- Sonya Huang, Pat Grady, Generative AI: A Creative New World, Sequoia Capital
This is a bold statement from Sequoia is something to think about to say the least. It's an exciting time to be in the Generative AI space!
(Part 3)
3. Dataset: A dataset is a data object that is many rows by many columns. In the Basejump UI, datasets are shown with a purple gradient and a table icon.
This is how we organize data in Basejump to help users manage their information. More information in our docs: lnkd.in/ga4JXXEF
3. Dataset: A dataset is a data object that is many rows by many columns. In the Basejump UI, datasets are shown with a purple gradient and a table icon.
This is how we organize data in Basejump to help users manage their information. More information in our docs: lnkd.in/ga4JXXEF
LinkedIn
This link will take you to a page that’s not on LinkedIn
lnkd.in
February 21, 2025 at 6:51 PM
(Part 3)
3. Dataset: A dataset is a data object that is many rows by many columns. In the Basejump UI, datasets are shown with a purple gradient and a table icon.
This is how we organize data in Basejump to help users manage their information. More information in our docs: lnkd.in/ga4JXXEF
3. Dataset: A dataset is a data object that is many rows by many columns. In the Basejump UI, datasets are shown with a purple gradient and a table icon.
This is how we organize data in Basejump to help users manage their information. More information in our docs: lnkd.in/ga4JXXEF
(Part 2)
2. Record: A record is a data object that is 1 row by many columns. In the Basejump UI, records are shown with an orange gradient and an ID icon
2. Record: A record is a data object that is 1 row by many columns. In the Basejump UI, records are shown with an orange gradient and an ID icon
LinkedIn
This link will take you to a page that’s not on LinkedIn
lnkd.in
February 21, 2025 at 6:50 PM
(Part 2)
2. Record: A record is a data object that is 1 row by many columns. In the Basejump UI, records are shown with an orange gradient and an ID icon
2. Record: A record is a data object that is 1 row by many columns. In the Basejump UI, records are shown with an orange gradient and an ID icon
Part 2
Some things to consider:
- What is the top K to keep from recent chats vs retrieval?
- Which vector database do you use?
- What do you include in the memory for the AI vs. what is shown to the user? - Do you add additional metadata to chat history?
Some things to consider:
- What is the top K to keep from recent chats vs retrieval?
- Which vector database do you use?
- What do you include in the memory for the AI vs. what is shown to the user? - Do you add additional metadata to chat history?
LinkedIn
This link will take you to a page that’s not on LinkedIn
lnkd.in
February 21, 2025 at 6:47 PM
Part 2
Some things to consider:
- What is the top K to keep from recent chats vs retrieval?
- Which vector database do you use?
- What do you include in the memory for the AI vs. what is shown to the user? - Do you add additional metadata to chat history?
Some things to consider:
- What is the top K to keep from recent chats vs retrieval?
- Which vector database do you use?
- What do you include in the memory for the AI vs. what is shown to the user? - Do you add additional metadata to chat history?
