We are making a science understanding system, starting with market research for scientific sales.

Selling to biotech companies requires understanding their science: combing through research papers, redlining patents, screening thousands of scientist LinkedIn profiles, etc for each prospective customer. This work often requires a PhD salesperson to spend up to 30+% of their time manually doing these intellectually-consuming but manual workflows.

Raycaster uses AI to save them time and land bigger deals faster.

Raycaster is building research agents for science, starting with scientific sales. 

The backdrop

• Selling to biotech companies requires understanding their science: combing through research papers, redlining patents, screening thousands of scientist LinkedIn profiles, etc for each prospective customer
• This work often requires a PhD salesperson to spend up to 30+% of their time manually doing these intellectually-consuming but manual workflows
• As a result, the pipeline for million-dollar deals is bottlenecked at the research point, even though AI can help make it dramatically easier

The company

• Levi — Built enterprise AI systems @ Uber, Zoom, Stanford
• Anthony —  Did engineering @ Apple, Meta, Snap
• Completed YC’s Fall 24 batch in December
• Featured on TechCrunch as a top company in the batch

Why?

• Our goal is not to compete in sales tech. We're going after scientific superintelligence, starting with the highest-leverage team - sales.

We want to work with someone who is

• A founder or are a former founder (CTO)
• Passionate for science 
• A little autistic
• Able to start at least part-time soon
• [nice to have] Dislikes YC as a concept
• [nice to have] Cracked at the ‘modern’ web stack (NextJS, Supabase, Vercel, serverless functions)

You can apply here or just text +1 510-930-8618 (iMessage) with a bit about yourself and we’ll find a time to chat.

• AI wise— we don't rely on any one framework much, but DSPy is going to play a bigger role in our future.
• Web wise— ‘modern’ stack (NextJS, Supabase, Clerk, Vercel, Tanstack tables, serverless functions).
• We carry over some code culture from Meta in terms of monorepo and ship-friendliness