Senior Software Engineer II – Agentic Intelligence

This is a fully remote position, open to applicants in United States.

📋 Description

• Create and implement production-ready agents. Develop agents that can investigate, reason, and take action based on live observability data within Canvas. These agents need to be reliable for engineers, especially in high-pressure scenarios, including during incidents. Transform initial rough prototypes into robust solutions that can handle production traffic.

• Take full ownership of the agent development; support the entire product lifecycle. Define, build, deploy, and maintain the agents, including evaluations that determine their effectiveness and improvements. This position is primarily agent-focused but also involves elements of full-stack development.

• Develop agents that are unique to Honeycomb. Utilize a data store capable of returning high-cardinality queries in seconds to analyze signals that conventional backends cannot support at this level of detail, allowing for service correlation, deep dives into individual traces, and comparisons before and after deployments.

• Broaden the capabilities and determine future directions. Introduce new features into Canvas, the MCP server, and enhance Canvas Skills memory, spatial awareness, and accelerate the Bedrock loop, while advocating for subsequent developments with functional code. Differentiate between genuine signals and hype in a landscape filled with both.

• Establish the criteria for what constitutes "good" agents in this context. Set measurable standards based on actual evaluations, ensuring maintainability while being transparent about their limitations.

⛳️ Requirements

• Experience in AI and agent engineering. You have successfully launched LLM-based systems that users depended on in production environments, not just demos or fine-tuned models in research settings. You understand the failure points of agent systems and how to design solutions that mitigate these issues.

• Comprehensive ownership of projects. In a small team, there are no handoff delays. You can advance a project from a rough prototype to a production-ready solution independently.

• Up-to-date insights, not merely historical knowledge. You possess well-informed perspectives on recent shifts in agent design over the past six to twelve months that would influence your current development approach.

• In-depth knowledge of agent architecture. You grasp how a fast, high-cardinality data store alters the reasoning capabilities of an agent and how to design effectively for that environment.

• Strong product insight. You can assess the current functionalities of the agent layer and identify its future enhancements, supporting your proposals with prototypes rather than presentations.

• Background in observability or developer tools. Engineers are your primary users; familiarity with this domain will accelerate your onboarding and enhance the quality of the work.

• Experience with evaluation frameworks, agent tooling, RAG, and prompt engineering.

🏝️ Benefits

• A stake in our success - generous equity with an employee-friendly stock program.

• Pay transparency based on established levels relative to experience, rather than negotiation skills.

• Unlimited PTO to allow for time to recharge.

• A distributed-first culture and mindset (we truly mean it!).

• Stipends for home office setup, co-working spaces, and internet costs.

• Comprehensive benefits coverage for employees, with additional options for dependents.

• Up to 16 weeks of paid parental leave, regardless of the path to parenthood.

• Annual professional development allowance.

• And much more...