Beta Update July #1 - Vector Store, Object Store, Podcast and much more!

Another update from the Flight Deck. This time with exciting updates on Vector Stores, Object Stores, upcoming events, the first podcast release, a new wave of beta sign-ups, and much more, so put on those aviators and keep reading!

Vector Database

You have asked, and we deliver; we have added support to the SDK for the Seaplane-managed vector store. You can read more about it in our documentation here. Just like any other task resource, Seaplane handles all the infrastructure complexity and leaves the data science up to you!

With the vector store, we now enable use cases such as in-context learning, powerful recommender systems, better document search, and much more. Vector DB is now available to select Seaplane beta users. If you want access today, reply to this email, and we will get you boarded as soon as possible.

Object Store

We have added support to the SDK for the Seaplane-managed object-store. The object store can store any object up to 8MB in size. The data in the object store is decoupled from the application life cycle, meaning you can persistently store data. Deleting your app or task does not affect the data in the object store. You can read more about it in our documentation.

With the object store, you can build more powerful workflows requiring persistent data. The object store is now available to select Seaplane beta testers. Want access today, let us know by replying to this email.

PDF Summary Tutorial

In one of our previous updates, we included a video deploying a PDF summary pipeline. This pipeline is now available as a tutorial here; you can take a look at the code in our demo Github Repository here. Keep an eye on this repository for future demos, tutorials, and open-source applications including.

• LLM-powered resume search. Imagine loading hundreds of resumes into this demo app. Now you can ask the LLM to show you everyone with experience in a,b,c...x,y,z
• LLM-powered sales call search. A demo app that takes in sales calls recordings and makes them available through a chat GPT-style interface.
• Internal knowledgebase chatbot. Many companies have hundreds, if not thousands, of pages of internal knowledgebase. Over time they become impossible to navigate. Not anymore! Simply ask your question to a chatbot and get a direct answer.

First Podcast Released

We have released the first podcast in our new series, “Scale Up AI,” where we talk to industry experts about scaling and building AI applications, Teams, and Companies. In our first edition, we invited Tom Rogers, who, among other things, helped build the first Machine Learning team at Lifion by ADP. You can now listen to it on Spotify and soon on Apple Podcasts.

Be sure to subscribe as we have exciting new episodes coming up with:

• Kathy Copic - Data Strategy & Hiring Consultant at Fieldwork about building and scaling data teams
• Debarag Banerjee - Head of AI at MachinAnimus - Cover his paper: AI Failures: A review of Underlying Issues
• George Williams - Generative AI researcher at Ansible AI consulting - About, well you guessed it, generative AI, the implications on the industry, and how to regulate it.

Silicon Valley AI Meetup

We are attending the first edition of the Silicon Valley AI Meet-Up and will be giving a live demo of the Seaplane platform. You can sign up here. We have a special offer for anyone attending, so be sure to join us if you can!

Wave Two Beta Onboarding

We started wave two of the beta onboarding and are adding new users weekly. Not yet signed up? Claim your spot here.

That was it for this edition, but keep your eyes peeled on the horizon for our next update with our first customer use case, new SDK features, podcast, and much more!

‍