At Sierra, we’re creating a platform to help businesses build better, more human customer experiences with AI. We are primarily an in-person company based in San Francisco, with growing offices in Atlanta, New York, London, Paris, Madrid, Munich, Singapore, Japan, and Sydney.

Agent Architecture: What primitives do we need to build agents that are steerable and verifiable, but also conversational and empathetic? How do we future-proof this as LLMs evolve?

Retrieval: How do we ground answers in a customer’s knowledge base? How do we use retrieved context conversationally, handling cases where the answer is unclear or needs clarification from the user?

Evals: How do we measure an agent’s quality? How do empower our customers to improve agent quality?

Voice: How do we make our chat agents “just work” over the phone? How do we deliver lifelike conversations at low latency?

Simulation & Benchmarking: How can we craft a simulation platform to test AI agents against every real-world scenario imaginable? (See 𝜏-Bench https://sierra.ai/blog/benchmarking-ai-agents)

Content Management: How do we create intuitive, no-code tools that allow anyone to guide and test AI agents?

Agent Development Lifecycle: How do we adapt traditional software development methodologies to accommodate AI agents’ non-deterministic behavior, natural language interactions, and reliance on large language models? (See Sierra’s ADLC https://sierra.ai/blog/agent-development-life-cycle)

Generative Agent Development: How do we accelerate agent creation using generative tools like Cursor and Claude Code? Can we build self-improving systems based on real-world interactions, customer-driven feedback, and self-play?