AI Assistant for Your Company's Private Data

Cubie is an easy, lightweight solution that lets your customers ask analytical questions in natural language about your large, murky datasets  
See Cubie in action with sample data from QuickBooks
What's your doc format?

What We Do


ChatGPT has been a thing for a while now.  It's an extremely useful tool for summarizing texts and answering questions about topics in the public domain.  But what about questions that require analyzing a vast amount of private data?  

Large organizations generate an enormous amount of information they want to keep private (e.g. sales figures, research findings) or are legally required to keep private (e.g. customer data).  Datacakes specializes in creating custom, wholly-owned systems where an AI (like GPT) can analyze your structured data, without the need to prepare or export it.

For a demo of how we've set up a custom platform to work with private data, check out CUBIE, our AI Assistant connected to QuickBooks.
To get an answer from ChatGPT on a topic that it hasn't already been trained on, you're expected to give it the source data it needs for the answer.  For example, when you want it to summarize a document, you have to copy/paste or upload the document.  

In constrast, our approach asks the AI to create instructions on how to answer your question, takes those instructions and prepares (quality-check) them to be executed on your data in your own safe space.  This way, your data can stay where it is, and no export is needed.
For some types of requests, giving some data to the LLM (Large Language Model, the tech behind ChatGPT) is unavoidable. In these cases, it may be worthwhile to host a custom LLM, where you control every aspect of its operation.

ChatGPT is the most well-known LLM service, but the field has exploded with open source competitors that we can help you configure, and further train, as an in-house alternative. Many popular cloud platforms (e.g. AWS, Google Cloud, Databricks, Cloudera) have recently introduced support for these custom LLMs, as their popularity grows.


Hallucinations are one of the major risk points of current LLM technology. Sometimes, an AI will make up people, events, or data that do not actually exist in order to answer a question it was asked. This is particularly difficult to deal with, because the LLM is good at making its answer sound like fact.

Hallucinations are made worse by an over-reliance on prompt engineering. It is tempting for the human user to load their question prompt full of requirements, to ensure a certain kind of response. But there is no guarantee that the AI will meet each of those requirements, and its unpredictability may even increase with an overloaded input.

For our custom solutions, we've developed a proprietary approach that combines LLM thoughts with more classical, deterministic functions. We've found that this reduces the variability of the overall response and allows the handling of more complex questions. (Our research in this area is ongoing.)

Who We Are

Our founder’s backgrounds comprise decades of experience in both scrappy startups and senior management at Fortune 500 companies.

Tim Eller

Tim is an AI innovator who is hell-bent on helping businesses make faster and better decisions with data. After originating Lyft's dynamic pricing algorithms in 2014, he refactored...


Rob Sarvis

Rob is a polymath whose education and professional careers have allowed him to indulge in his principal joys: learning about the world, and enabling progress through marvelous tech...


Paul Sun

Paul is a creative business leader and occasional hacker whose career has spanned finance and technology. He led the Capital One Innovation Lab in San Francisco, where his organization...


Frequently Asked Questions

Get In Touch

Let us know if you need a customized solution that Cubie doesn't cover
How would you describe your organization?
Thanks for your interest, we'll be in touch shortly!
Oops! Something went wrong, please try submitting again.