Member of Technical Staff (Research Engineering)

Code Evaluation

Descripción

Job type: Full-time.
Location: Remote.
We are seeking a Research Engineer to operate at the frontier of reinforcement learning (RL), developing novel environments, training pipelines, and evaluation systems that advance the capabilities of modern AI models.
This role sits at the intersection of research and production, translating experimental ideas into scalable, high-performance systems.

Architect self-contained RL environments that capture complex, real-world tasks, including reward functions, verifiers, and evaluation logic.
Design and episode pipelines and multi-component training processes (MCPs) to support reproducible experimentation.
Build automated data generation systems, leveraging synthetic data to accelerate training cycles without compromising quality.
Build and integrate AI-driven evaluation and quality assurance systems for automated grading, validation, and feedback loops.
Fine-tune and optimize open-source RL models using internally generated datasets and custom training strategies.
Establish benchmarking frameworks to measure model capability, robustness, and data quality across tasks.
Contribute to the release and analysis of evaluations on internal and external benchmark platforms (e.g., the company benchmarks).

Deep background in reinforcement learning, including environment design and training dynamics.
Strong track record of building and scaling RL systems, pipelines, or experimentation frameworks.
Proficient in automation and data generation, including synthetic data pipelines.
Familiar with automated evaluation systems, model validation, and quality assurance workflows.
Experienced in fine-tuning and evaluating open-source ML models.
Clear, concise communicator with strong technical writing skills.
Comfortable operating in fast-paced, research-driven, and highly collaborative environments.

publishing benchmarks, evaluations, or research artifacts.
Familiarity with evaluation ecosystems (e.g., the company benchmarks or similar frameworks).
Background in scalable infrastructure for large- RL experimentation.