Senior Backend Engineer – Agents (USA Only - 100% Remote)
About The Position
The Agents team owns the platform underneath our first agent, Chloe, which was shipped in 2026. This includes the agent core, the evaluation and observability layer, the MCP surface that external agents (Claude, ChatGPT, ElevenLabs, n8n) operate through, and the orchestration that ties it all together. The team is currently focused on three major streams: Voice Agents, the LLM-powered chat assistant, and Custom Agents. Many of the engineering challenges in this area are novel and are being solved in production with a user base of 11,000 paying customers. This role is open at multiple levels: Software Engineer, Senior, and Staff. Engineers may be moved between teams as work shifts, offering opportunities to work on various projects over time.
Requirements
- Seasoned engineer with experience in Python, and potentially Go, Rust, or TypeScript.
- Experience shipping meaningful, impactful agentic features to users.
- Familiarity with retrieval, evals, tool design, and context engineering.
- Experience using AI tools in a daily workflow to improve speed, code quality, and understanding of codebases.
- A builder mentality, prioritizing shipping a functional v1 over extensive abstractions.
- Experience debugging incidents, owning critical systems, or carrying a pager for systems with customer impact.
- Familiarity with reading research papers and assessing their relevance to practical problems, particularly in areas like retrieval, RAG, RLHF, or fine-tuning.
- Comfortable working with non-deterministic systems and probabilistic outputs.
Nice To Haves
- Experience with Python, Go, Rust, or TypeScript.
- Experience with retrieval, evals, tool design, context engineering.
- Experience using AI tools in a daily workflow.
- Experience debugging incidents, owning critical systems, or carrying a pager.
- Familiarity with reading research papers and assessing their relevance to practical problems, particularly in areas like retrieval, RAG, RLHF, or fine-tuning.
- Comfortable with non-determinism.
Responsibilities
- Ship code execution for the assistant, enabling it to decide when writing code (e.g., generating charts and tables in Python, using user-defined tools, calling external APIs) is more effective than non-deterministic answers.
- Build the evaluation and observability layer to measure agent improvement, using evals and tracing (LangFuse and internal tooling) as a prerequisite for shipping.
- Advance the generated UI by having the backend determine the user-facing presentation (table, chart, widget) and render it in the assistant, with plans for full-screen experiences and stored artifacts.
- Develop Custom Agents from prototype to General Availability, focusing on event-driven agents that act in real-time within the CRM.
- Integrate deterministic and non-deterministic systems, fusing the Workflows engine with agentic steps to ensure reliability where needed and intelligence where beneficial.
- Address the edge cases that ensure agent trustworthiness, such as handling paused semantics, recovery, and managing agent calls when AI credits are depleted.
- Select the appropriate AI model for specific tasks, considering multiple providers, continuous testing, and cost-aware routing to avoid using expensive models for simple tasks.
Benefits
- Opportunity to work on cutting-edge AI and agentic systems.
- Potential to work on multiple projects and teams over time.
- Company retreats (e.g., Milan, Dusseldorf, Germany).
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
Education Level
No Education Listed
