Backend Engineer

LiteLLM•San Francisco, CA

15h

About The Position

LiteLLM is the world's most popular AI Gateway, trusted by top companies like Adobe, Netflix, and NASA. Our platform empowers developers by providing secure, reliable access to LLMs and adjacent services, and we're looking for a Senior Backend Engineer to help us build rock-solid guardrails and observability tooling at scale. About The Role You’ll focus on owning our guardrails and logging world-class. You will be in charge of the backend code that ensures all guardrail calls are consistently logged, errors are surfaced to users (not silently swallowed), and our observability instrumentation works for real-world, high-volume traffic. Your attention to detail in areas like latency metrics, logging traceability, and backend guardrail registration will directly impact user trust in our security and compliance features.

Requirements

Bachelor’s or Master’s in Computer Science or related field
1+ years of experience with Python and backend frameworks (e.g. FastAPI, Flask)
Understanding of logging best practices, error handling, and secure backend development
Exposure to monitoring, logging, or metrics platforms (Datadog, Splunk, Prometheus, OpenTelemetry)
Familiarity with database integration and troubleshooting (PostgreSQL, Redis, etc.)
Driven to deliver high-quality backend code with strong guardrails, auditing, and debugging capabilities
Eagerness to tackle hard bugs and ensure system transparency for end users

Responsibilities

Build and scale our product, ensuring performance, reliability, and continuous improvement.
Ensure all guardrail and policy enforcement calls (e.g., applyguardrail ) are properly logged and traceable through our SpendLogs and relevant database tables
Build and design CPU-level guardrails to cover common attacks on LLM API's / MCP servers / Agents
Identify and fix areas where silent failures occur in guardrail creation, registration, and policy application—ensuring robust error handling and transparency to end users
Work with observability integrations, including Datadog, Splunk, Prometheus, and OpenTelemetry, to maintain accurate, configurable, and usable monitoring and logging for backend systems
Enhance observability integrations to work for 1B+ requests/mo., with minimal latency overhead and no memory leaks (e.g. due to cardinality of Prometheus metrics)
Collaborate cross-functionally on backend engineering priorities (performance, reliability, security)