Staff Research Engineer – RRI Harness

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

📋 Description

• Develop and enhance the Python multi-agent runtime, focusing on orchestration, tool contracts, context and memory management, compaction, and model adapters across various LLM providers.

• Design and conduct experiments related to agent prompting, tool design, context and memory strategies, and model selection.

• Implement changes that significantly improve RRI results.

• Provide a fixed evaluation set, a scorecard with measured noise bands, evaluations for each change, and component ablations.

• Engineer resilience against GPU failures, node replacements, rolling upgrades, provider errors, and credit limits.

• Create event streams, perform cross-session analyses, manage per-agent cost and token accounting, and develop MCP administration tools.

• Identify and minimize agent idling, looping, and excessive token utilization.

• Design interfaces for the Go control plane and GPU proxy owners.

• Collaborate with applied scientists and forward-deployed engineers to troubleshoot and resolve client issues in production.

• Take ownership of the core runtime for Sequen's autonomous research engine.

⛳️ Requirements

• Over 7 years of experience in developing production software.

• Proficient in Python.

• Experience with distributed or long-running systems.

• Familiarity with LLMs in agentic loops, including tool calling, prompt and context management, streaming, and retries.

• Capable of debugging across processes, containers, and network boundaries.

• Experience with Kubernetes and GPU nodes.

• Ability to interpret a training script, comprehend a ranking metric, and differentiate actual regressions from noise.

• Applied research experience with ML or LLM experiments, including baselines and ablations.

• Experience with evaluation infrastructure, test harnesses, benchmarks, experiment tracking, or ML CI.

• Basic knowledge of Go.

• Familiarity with on-prem, BYOC, or air-gapped environments and their security constraints.

• Experience with large-scale PyTorch training, GPU scheduling, or ML platforms.

• Contributions to agent frameworks, evaluation harnesses, or ML tooling are considered a plus.

🏝️ Benefits

• Unlimited paid time off.

• Flexible hybrid/remote work configurations.

• A highly collaborative, world-class engineering culture.

• Equity.