Decagon
Staff Software Engineer, Agent Platform
$200K–$400K
San Francisco · Hybrid · Staff · Full-time · Posted Dec 1, 2025
About the job
Designs and operates the distributed runtime, orchestration, and experimentation systems that make Decagon’s AI agents reliable, safe, and effective in production.
What you'll do
- Build agent harnesses, runtime systems, and multi-model orchestration, including routing, planning, and tool invocation.
- Improve production latency, reliability, safety, observability, and correctness; investigate failures and drive iterative improvements.
- Build online experimentation and contribute to offline evaluation, testing, simulation, and real-time voice systems.
What they ask for
- Strong experience building production distributed systems or backend platforms.
- Experience owning systems end-to-end, from design through production and iteration.
- Familiarity with experimentation, evaluation, or data-driven improvement loops.
- Track record of improving reliability, performance, and observability; ability to debug complex systems.
Nice to have
- Experience with agent harnesses, orchestration layers, or execution frameworks.
- Systems-level thinking about control planes, feedback loops, and optimization.
Pay
$200K – $400K USD base salary
Tools
A/B testing
AI focus
Build and evolve agent runtime and orchestration systems, coordinate models and tools, enforce safety, and improve agent behavior through evaluation and experimentation.
Model-written summary · the original posting is authoritative