April Update - 30+ models on Seaplane.io

It's been a while since you last heard from us, and we're excited to bring you the latest updates through our newsletter. We're committed to keeping you in the loop more regularly moving forward. Here's a quick rundown of what's inside:

Quick Summary

• Dive into the excitement of our March Model Madness event – your vote counts!
• Join us for an upcoming live event where we'll demonstrate building a LLM-powered voice assistant in under an hour. Details here.
• Explore our recent demos to see Seaplane in action: from simplifying model access, to creating an AI-enhanced Shopify store, and developing a RAG application with Seaplane.

March Model Madness

You're probably familiar with the excitement of March Madness, but have you caught wind of March Model Madness yet? This year, Seaplane and the MLOps Community are upping the ante with the most anticipated LLM showdown.

In this event over 20 cutting-edge models are going head-to-head in a blind competition. It's up to our community to decide which models stand out by voting on the most impressive responses.

The competition is fierce across three categories: chat, instruct, and code. Only the top performers will advance, battling it out to earn the title of the ultimate LLM in each category.

But that's not all—we're also spotlighting the artistic side of AI. Four leading image generation models are fighting for the crown of 2024's best image model. Your vote not only helps us pick the best model in each category but also enters you into a raffle to win great prizes.

Ready to have your say? Cast your vote at marchmodelmadness.com and be part of deciding this year's leading models in AI.

What's New in the Seaplane SDK

Our recent updates are bringing a ton of new powerful features to the platform. For a deep dive into all our latest improvements, don't forget to check out our technical documentation.

Here’s a sneak peek at what’s new:

A Universe of Models

The Seaplane AI model collection has grown exponentially! We've introduced over 30 generative AI models, making our platform richer and more versatile than ever. Highlights include:

• Llama Series: A comprehensive across chat, code and instruct models, available in 7b, 13b, 34b, and 70b configurations.
• Mistral: Versions 0.1 and 0.2 of this 7b model promise incredible speeds with less than 2 seconds of inference time on average prompts while not budging on quality of output.
• Zephyr 7b: Elevate your AI applications with Zephyr's capabilities.
• Stable Diffusion XL: Push the boundaries of image generation with this advanced model.
• Styletts2: A super fast test to speech model that can rival the likes of OpenAI.
• Claude3 Sonnet and Haiku: Anthropic recently released their new V3 models and they are already available on Seaplane. Said to beat OpenAI’s GPT-4 in many tasks.
• Gemini-pro: A powerhouse for a range of generative tasks.
• OpenAI GPT Series: Including, GPT 3.5, GPT 4, and the larger GPT 4-32K.

Accessing these models is a breeze with Substation, our model hub. You can access any of them by changing a single line of code.

Advanced Input, Output, and Routing in DAGs

We've significantly upgraded our Directed Acyclic Graphs (DAGs) infrastructure to support multiple inputs and outputs. This enhancement simplifies the execution of complex workflows, such as conducting A/B testing or managing parallel data processing paths with ease.

Furthermore, we're excited to introduce conditional routing. This new feature empowers you to direct data flow within your DAGs based on specific criteria, optimizing task execution based on real-time inputs. Learn more about conditional routing in your projects.

Introducing Flightdeck Visualizations

Responding to your feedback, we acknowledge the challenge of managing and debugging large-scale DAGs. To address this, we're proud to unveil our latest feature: Flightdeck DAG Visualizations. This tool enriches Flightdeck by offering an intuitive visual representations of your DAGs, complete with task-level log visibility. Debugging and understanding complex DAGs has never been easier or more efficient.

Demos & Live Events

This quarter, we've been busy at Seaplane creating some truly innovative demos to showcase the capabilities of our platform. Here’s a glimpse of what we’ve been up to:

Personalized AI Shopify Store

A Shopify store that tailors product offerings based on real-time chat interactions with users. We made it happen! This personalized AI-powered shopping experience is a testament to the customizability and power of Seaplane. Watch the demo here or dive into the nuts and bolts of how we built it in our technical deep dive blog post. Did you know that this demo uses Seaplane operated GPUs for incredibly low latency?!

LLM-Powered, RAG-Enabled Voice Assistant

Step aside, Siri and Alexa. We've developed a voice assistant that leverages Large Language Models (LLM) and Retrieval-Augmented Generation (RAG) for unparalleled performance. The full video demo will be available shortly, but you can get ahead by exploring our technical breakdown now.

Live Coding Session: Build With Us

Don't just watch—participate! Join us for an exciting live coding session where we'll build an LLM-powered, RAG-enabled voice assistant from scratch in just 60 minutes. This is a rare opportunity to see the magic happen in real time and learn how to apply these technologies to your projects. Sign up on LinkedIn to secure your spot.

Join Our Team as an SDK Engineer (100% Remote)

We're on the hunt for an SDK Engineer to expand our fully remote team. If you're passionate about crafting exceptional Python applications and have a knack for building user-friendly SDKs, we're looking for you!

What You'll Bring to the Table:

• Python Mastery: 5+ years of experience in developing Python applications, with a strong portfolio in customer-facing Python SDKs.
• Stream Processing Expertise: At least 3 years of working with stream processing technologies like Kafka, Apache Spark Streaming, or NATS.
• Data Storage Savvy: Knowledgeable in modern data storage solutions, including distributed SQL, Vector Stores (e.g., Qdrant, Pinecone), and object stores.

You’ll Capture Our Heart If:

• Startup Spirit: You’ve thrived in the dynamic environment of a small to medium-sized startup, valuing collaboration and agility.
• Data Science Acumen: Your background includes working in the data science field, with familiarity in tools, packages, and trends in Data Science and AI, including LLMs, Langchain, and Generative AI.
• Remote Work Pro: You have proven experience thriving in a 100% remote work setting where clear, concise written communication is paramount.
• Customer-Centric Development: You have a track record of developing data science tools or experiences that customers adore.
• Project Management Prowess: You’re adept at steering projects using management tools like Linear.

Join us in our mission to revolutionize data science and AI technologies. Apply now to be part of a team where innovation, collaboration, and customer satisfaction are at the core of everything we do.

‍