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

National institutions, public-sector organizations, and regulated industries need capable models and agents that operate within strict boundaries around data, deployment, access, auditability, and human authority. They also need the ability to improve those systems on their own workflows rather than depend indefinitely on a closed external model. You will build post-training and agent systems in public-sector and regulated-industry settings.

Responsibilities

Applied Research at Prime

We turn a workflow into tasks, environments, evaluations, and reward signals, run post-training experiments against real baselines, and deploy what works. When a result looks good, we find out whether the gain came from the model or from the data, harness, inference, or infrastructure. The reusable parts become shared research, open-source systems, and product capabilities.

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

Role scope

The work happens inside partner constraints, sometimes on their infrastructure and under their access rules.

Representative early problems

Build a private improvement loop that turns sensitive traces into environments, training signal, evaluations, and deployable model updates without crossing the operator's data boundary.

Train a tool-using domain agent whose actions are permissioned, traceable, recoverable, and evaluated against real procedures rather than generic benchmarks.

Develop high-assurance evaluations for reliability, calibration, robustness, human escalation, and compliance with domain constraints under adversarial or changing conditions.

What you will own

Improvement loops that run end to end inside the operator's boundary, from trace selection to deployed update.

Agents with explicit permissions, evidence trails, bounded tools, and human checkpoints.

Evaluations that test reliability, calibration, recovery, and compliance with real procedures.

Experiments on how privacy, access, and infrastructure constraints change learning, with methods that transfer across high-assurance settings.

Requirements

Relevant experience: RL/post-training, private ML systems, high-assurance evaluation, agent safety, regulated-domain ML, public-sector technical work, or deployment under strict data and infrastructure constraints.

You should want to build and evaluate these systems directly.

Hands-on experience building and debugging model-training, post-training, agent, evaluation, or closely adjacent experimental loops.