Staff Research Engineer: Autonomous AI Agent Runtime (Remote/Hybrid)

Openkrill; United States; Hybrid; USD 180,000 - 240,000; Full time

Benefits offered by this job

Unlimited PTO

Hybrid/Remote configurations

Collaborative culture

Job summary

Sequen is seeking a Staff Research Engineer to own the harness and contribute to applied research for the Recursive Ranking Intelligence (RRI) platform. You will run multi-day experiments on GPUs inside client clouds or Sequen-hosted environments, and build the Python runtime, tools, and adapters that light up long-running agent sessions.

You will design clean contracts with the Go control plane, partner with applied scientists, and deliver reliable, observable systems with a bias toward

Qualifications

• 7+ years building production software with Python and distributed systems.
• Experience with LLMs in agentic loops and scaling challenges.
• Comfort debugging across process, container, and network boundaries.
• Ability to own a system end to end with strong reliability focus.

Responsibilities

• Own the harness: Build and evolve the Python multi-agent runtime across runners, contracts, and adapters.
• Contribute to applied research: Design experiments and ship changes that improve results.
• Build eval infrastructure: Deliver fast eval sets and scorecards for every change.
• Engineer for resilience: Ensure multi-day sessions survive failures and upgrades.
• Deliver observability: Create event streams and cost accounting across sessions.
• Cut waste: Identify idle loops and token waste and fix them in code.
• Design clean contracts: Collaborate on interfaces with the Go control plane and GPU proxy.
• Support clients: Work with scientists and engineers to diagnose production issues.

Skills

Python

Distributed systems

LLM agenting

Kubernetes

Observability

Troubleshooting

Tools

Go

PyTorch

Docker

CI/CD

Job description

Sequen is seeking a Staff Research Engineer to own the harness and contribute to applied research for the Recursive Ranking Intelligence (RRI) platform. You will run multi-day experiments on GPUs inside client clouds or Sequen-hosted environments, and build the Python runtime, tools, and adapters that light up long-running agent sessions.

You will design clean contracts with the Go control plane, partner with applied scientists, and deliver reliable, observable systems with a bias toward