Applied Research - Scientific AI

OpenKrill; United States; Remote; USD 150,000 - 300,000; Full time

Benefits offered by this job

Cash compensation

Flexible work locations

Visa sponsorship

Professional development budget

Team offsites

Job summary

Prime Intellect is seeking researchers to build post-training agents for scientific research, including literature search, code execution, and tool use. You will operate at the frontier of autonomous research, collaborating with researchers and engineers to turn workflows into trainable tasks within a fixed compute budget.

You will design verifiers for correctness and reproducibility, develop RL methods for proposing and critiquing hypotheses, and ensure experiments are well-documented and

Qualifications

• Depth in RL, agents, or post-training systems.
• Experience building training/evaluation loops.
• Strong Python and tooling for reproducibility.
• Ability to collaborate with researchers and engineers.
• Comfort with ambiguous problems and stakeholder updates.

Responsibilities

• Build long-horizon environments for literature search, code execution, and experiments within a fixed compute budget.
• Turn research workflows, data, and criteria into trainable tasks, evaluations, and rewards.
• Design verifiers for correctness, novelty, and reproducibility; check for leakage and shortcuts.
• Link programmatic checks, simulations, domain models, and expert feedback to training; handle noisy or delayed results.
• Develop RL and multi-agent methods to propose, test, and critique hypotheses; compare with baselines.
• Maintain reproducibility of code, data, logs, and configurations; rerun experiments and document results.

Skills

RL expertise

Post-training systems

Python

Experiment design

Autonomous research

Tools

Domain tools

Job description

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

You will build and post-train agents for scientific research: searching literature, writing and running code, using domain tools, and choosing the next experiment. The work spans autonomous research and computational science, with environments and evaluations built around research partners' tools, data, and experimental feedback.

RESPONSIBILITIES

Every Applied Research role is externally facing. Most researchers work directly with customers, users, or research partners while maintaining a substantive research agenda.

Prime's autonomous-research study https://www.primeintellect.ai/blog/measuring-autonomous-research used multi-day nanoGPT optimization tasks to examine how agents measure noise, revisit earlier experiments, and verify improvements. You will build similar experimental loops for scientific tools and datasets.

• Build long-horizon environments where agents search literature, query scientific databases, run code and domain tools, and choose experiments within a fixed compute or experimental budget.
• Turn research partners' workflows, datasets, and outcome criteria into trainable tasks, evaluations, and reward signals.
• Design verifiers for correctness, novelty, and reproducibility. Test whether agents use the underlying evidence, with checks for label leakage, lookup shortcuts, and unsupported conclusions.
• Connect programmatic checks, simulations, domain models, and expert feedback to training. Account for noisy or delayed results, calibrate confidence estimates, and define where an experiment needs human review.
• Develop RL and multi-agent methods for proposing, testing, and critiquing hypotheses. Compare them with strong single-agent baselines at equal compute and evaluate on held-out problems.
• Keep code, data, logs, and experiment configurations reproducible. Re-run promising results, measure noise, and use ablations to separate model improvements from changes to the data, harness, or inference setup. Release reusable environments, evaluations, and research code.

REQUIREMENTS

• We are looking first for depth in RL, agents, or post-training systems and the ability to acquire the relevant scientific context.
• Prior work in autonomous research, ML for science, scientific computing, or reproducible computational research is useful but not required.
• Hands-on experience building and debugging model-training, post-training, agent, evaluation, or closely adjacent experimental loops.
• Strong experimental judgment: meaningful baselines, careful trace inspection, clear failure attribution, and a willingness to change direction when the evidence changes.
• Strong Python and the ability to build artifacts that other researchers and engineers can run, inspect, and extend.
• Comfort owning ambiguous problems and keeping partners and teammates current as you resolve them.

Evidence of exceptional work

We value publications, open-source systems, benchmarks, and shipped products as evidence of exceptional work. Show us something difficult you built, what failed, and how the evidence changed your approach.

BENEFITS AND PERKS

• Cash Compensation Range of $150-300k + equity incentives
• Flexible Work (San Francisco, New York, London, Berlin, or remote)
• Visa Sponsorship & relocation support
• Professional Development budget
• Team Offsites & conference attendance

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.