Research Engineer - Data and Evaluation

Prime Intellect, Inc.; San Francisco (CA); Hybride; USD 150 000 - 350 000; Plein temps

Avantages offerts par ce poste

Cash compensation

Remote or SF office

Visa sponsorship

Relocation assistance

Team off-sites

Learning opportunities

Résumé du poste

Prime Intellect, Inc. is building an open frontier AI stack and seeks a role to design and build evaluations and environments for RL and long-horizon agents.

You will work directly with researchers to investigate model failures, from task implementation to scoring criteria. You will design and scale data pipelines, verify reward designs, and help turn complex data into rigorous measurements that drive real research impact and product-ready capabilities.

Qualifications

• Strong software engineering skills in Python and experience building data pipelines, evaluation systems, or ML workflows used by others.
• Hands-on experience with language models or agents and understanding impact of training data on behavior.
• Ability to inspect complex data and model outputs, identify failure modes, and turn an ambiguous quality question into a testable measurement.
• A record of owning technical projects through experimentation, implementation, and validation, while collaborating closely with researchers and engineers.
• Hands-on experience building technical evaluations or agent environments, with the engineering depth to inspect model traces and debug harness and scoring code.

Responsabilités

• Design and build challenging evaluations for coding and long-horizon agents.
• Build executable environments in Verifiers, with realistic task setup and clear success criteria.
• Develop and test scoring logic and verifiers, investigating reward hacks and misleading measures of progress.
• Trace failures through model outputs and harness code to distinguish model limitations from problems in the environment or scoring.
• Build and scale RL data pipelines that generate and curate challenging coding and agentic tasks.
• Work with RL researchers to test whether changes to tasks and training data improve performance on held-out tasks.

Connaissances

Python

Data pipelines

Evaluation systems

ML workflows

Research collaboration

Outils

Verifiers

Prime-RL

Agent environments

Description du poste

Own Your Intelligence

Prime Intellect is building the open superintelligence stack: the infrastructure frontier AI labs build internally, made available to every ambitious AI team.

Our platform, Lab, unifies compute, environments, evaluations, secure sandboxes, high-performance training, and deployment into one full-stack system for post-training at frontier scale - from SFT and RL to tool use, agent workflows, and continuously improving production models. We are building open frontier AI: open-source models trained end to end for long-horizon tasks like autonomous research, and the full-stack platform our own research team uses to build them. The next generation of AI companies, enterprises, and research teams do not just need more GPUs. They need the ability to turn their own workflows, tools, data, and feedback loops into superintelligence they own.

Prime Intellect has raised $150M in total funding from Founders Fund, Radical Ventures, NVIDIA, and exceptional AI, infrastructure, and enterprise operators, including Andrej Karpathy, Dwarkesh Patel, and leaders and founders from Ramp, Perplexity, Harvey, Mercor, Zapier, Datadog, Cognition, OpenAI, Thinking Machines, Together AI, SemiAnalysis, LangChain, Browserbase, Cloudflare, Sierra, Databricks, Airbnb, OpenRouter, Standard Intelligence, Fleet, Core Auto, and more. We are looking for people who want to build at the intersection of frontier research, real infrastructure, and go-to-market for a category that does not fully exist yet.

Role Impact

Build the tasks and environments that train Prime Intellect’s models through RL, and the evaluations that show what they can actually do. You’ll work directly with researchers to investigate model failures, from how a task is implemented to how success is scored.

Responsibilities

• Design and build challenging evaluations for coding and long-horizon agents.

• Build executable environments in Verifiers, with realistic task setup and clear success criteria.

• Develop and test scoring logic and verifiers, investigating reward hacks and misleading measures of progress.

• Trace failures through model outputs and harness code to distinguish model limitations from problems in the environment or scoring.

• Build and scale RL data pipelines that generate and curate challenging coding and agentic tasks.

• Work with RL researchers to test whether changes to tasks and training data improve performance on held-out tasks.

Requirements

• Strong software engineering skills in Python and experience building data pipelines, evaluation systems, or ML workflows used by others.

• Hands-on experience working with language models or agents, with an understanding of how training data and evaluation choices affect model behavior.

• Ability to inspect complex data and model outputs, identify failure modes, and turn an ambiguous quality question into a testable measurement.

• A record of owning technical projects through experimentation, implementation, and validation, while collaborating closely with researchers and engineers.

• Hands-on experience building technical evaluations or agent environments, with the engineering depth to inspect model traces and debug harness and scoring code.

Relevant experience may include:

• Large-scale synthetic data generation, dataset curation, or distributed processing for model training.

• Evaluation methodology, executable coding tasks, agent environments, verifiers, or reward design.

• Experiment analysis, statistical testing, contamination detection, and reproducible research tooling.

Nice to have

• Experience with RL or post-training data, including task difficulty selection and analysis of rollout or reward distributions.

• Contributions to Verifiers, Prime-RL, or comparable environment and evaluation frameworks.

• Work on model-based grading, long-horizon coding agents, or diagnosing reward exploitation.

We welcome depth in data generation, environments and rewards, or evaluation systems. You don’t need to have worked across every area.

Benefits and Perks

• Cash Compensation Range of $150-350k, including equity incentives, aligning your success with the growth and impact of Prime Intellect.

• Flexible work arrangements, with the option to work remotely or in-person at our offices in San Francisco.

• Visa sponsorship and relocation assistance for international candidates.

• Quarterly team off-sites, hackathons, conferences and learning opportunities.

• Opportunity to work with a talented, hard-working and mission-driven team, united by a shared passion for leveraging technology to accelerate science and AI.

If you're excited about the opportunity to build the foundation for the future of decentralized AI and create a platform that empowers developers and researchers to push the boundaries of what's possible, we'd love to hear from you.