Vellum

At Vellum, our mission is to be the primary platform companies use to bring AI into production. AI is the biggest technology shift since the internet – the applications of it are everywhere! We expect the world to be dramatically different in the future as foundation models get better and better. We intend to be the platform layer that brings this amazing technology to life in production use cases.

We enable companies to bring real business value from AI / LLMs. The output generated by AI / LLMs is usually stochastic and bringing AI into production requires robust testing and choosing the right model for the job. We’ve seen companies spend months of engineering time building good tooling for this (time which could be spent on their core end user facing product) yet despite this, they struggle to get the best results out of these models. Our platform provides companies best in class tooling to use their own data in production AI applications.

We’re a YC W23 company, raised $10m, and are backed by top investors like Rebel Fund, Eastlink Capital, Arash Ferdowsi (co-founder of Dropbox), Dharmesh Shah (co-founder of Hubspot), Dan Scholnick (board member of Weights & Biases, Docker, New Relic), and Divya Bhat (former YC group partner, 2x CEO). We announced our fundraise on TechCrunch here: https://techcrunch.com/2023/07/11/prompt-engineering-startup-vellum-ai/

We've been growing at 15% MoM consistently and are looking to continue that trend for the next several quarters. This is an exciting time to join early on our journey!

P.S: Why the name Vellum?

• Vellum is a type of paper - the Magna Carta was written on Vellum. We intend to be the standard source of record in production
• You can’t spell Vellum without LLM 😉

Technical problems we're solving

There are so many interesting problems we need to solve: prompt chaining, fine-tuning infrastructure, open source models and multimodality, just to name a few. Ultimately, the goal is to bring engineering best practices to the fast moving world of AI/LLMs and ensure that our customers have the best models for the job trained on their own company specific data

Our stack

In case you’re curious:

• Google Cloud Platform - as our cloud provider
• Postgres - as our primary database
• Python - what we write our backend in
• Django - for our webservers
• Flask - for our microservices
• Typescript - what we write our frontend in
• React - as our frontend framework