About Paradigm

Our mission is to build a new type of workspace that lets every knowledge worker leverage the power of LLMs to scale immense amounts of research in a simple and intuitive UI. Paradigm lets users utilize its purpose-built infrastructure to orchestrate hundreds of thousands of agents to process millions of tokens per minute, compressing days of research into seconds.

Our team must confront and solve problems unique to parallelizing agents at this magnitude, not limited to: race conditions for 10^5 agents writing to a data model simultaneously, realtime transmission of agent outputs across all services to the client, optimizing for both write throughput and in-memory footprint for each agent.

What we’re looking for

• 2-5 Years of Experience as a Software Engineer.

• Consistent record an individual contributor to several medium to large projects.

• Consistent record as primary individual contributor to at least one medium scale project.

• Managed, designed and deployed high-traffic products in a production environment.

• Seen as a resource by peers who can help debug and answer questions.

• Evaluated when to import a new dependency or build internally

What can set you apart

• We want your skills, not your title. We will prioritize those with a strong, demonstrated skillset over those who have learned to effectively climb the corporate ladder.

• You know how to convey complex concepts clearly, can facilitate difficult conversations, and ensure everyone is on the same page.

• Prior experience with: React, TypeScript, Go, Postgres.

• Bonus if you're familiar with: Tanstack (Query, Router), Websockets, Zustand

What your day could look like

• Building the UI for agent research traces

• Codifying team-wide processes into skills to share

• Debugging connectivity between client and server states to improve consistency

• Defining models to standardize API contracts

• Triage issues and running ad-hoc RCA's to maintain product stability

What we believe

Our office is based in San Francisco. We prioritize candidates who will be available for in-person collaboration.