Kepler Ai
Agentic Systems Engineer
New York City. Posted 9 months ago
The role
You will build agentic infrastructure that powers an AI research platform. You will work on distributed execution frameworks, evaluation systems, context management, and ontology and provenance systems.
What you will do
- Build agent execution infrastructure including distributed systems for parallel agent orchestration
- Build evaluation systems to measure agent quality and catch regressions
- Optimize agent performance through context compression, prompt optimization, model routing, and latency reduction
- Build ontology and provenance systems to map concepts and trace outputs to sources
- Integrate AI into production with error handling, fallback mechanisms, and cost optimization
- Own systems end-to-end from design to production including services, database optimization, deployment, and monitoring
- Ship with production excellence using testing, monitoring, and deployment pipelines
Required qualifications
- 7+ years building production software
- Deep distributed-systems experience
- Experience designing systems that scale and handle complex workflows
- Experience building with LLMs, agent frameworks, or ML infrastructure
- Git workflows
- CI/CD
- Automated testing
- Observability
- Strong communicator who can discuss technical trade-offs clearly
Nice to have
- Large datasets
- ETL pipelines
- Knowledge graphs
- Semantic systems
Benefits
- 100% covered top-of-the-line medical, dental, and vision insurance for employees and their families
- HSA maxed by the company to the IRS limit
- Automatic coverage for life, AD&D, and disability insurance
- Daily lunch in office
- Unlimited PTO policy
- Development environment budget including MacBook Pro, monitors, ergonomic setup, and development tools
- Build anything budget for tools, libraries, datasets, or infrastructure
- Learning budget for conferences, courses, or programs
About the company
Kepler is an agent harness and infrastructure layer that makes AI model outputs reliable, traceable, and verifiable. The founding team includes former Palantir engineers and a repeat founder CEO.