Contribute to post-training and evaluation infrastructure for a novel multimodal AI system.
Fine-tune models using supervised and reinforcement learning approaches (e.g., SFT, RLAIF, RLVR).
Define custom metrics and develop tools to assess and mitigate AI-related risks in generative AI.
Sony AI is Sony's research organization using AI to unleash human creativity, working with gaming, entertainment, and music divisions. The Ethical Creative and Protective AI Lab operates as a globally distributed team across the U.S. and Canada, with some 900 million Sony devices in use worldwide.
Develop RL algorithms for reasoning, coding, and long-horizon tasks.
Design and run scaling experiments to improve training recipes across model sizes.
Extend RL training frameworks for asynchronous workflows and reliability.
Prime Intellect builds the open superintelligence stack, offering a platform that unifies compute, training, and deployment for frontier AI teams. With $150M in funding and a mission-driven culture, they focus on open-source AI research and real infrastructure.
Build the RL post-training stack end-to-end: rollout ingestion, reward computation, policy updates, and returning updated weights to the network.
Invent algorithms for asynchronous, high-latency, partially trusted generation, including staleness tolerance and off-policy corrections.
Ship the first decentralized post-trained models, building evals and taking the release from run to public artifact.
Pluralis Research is developing Protocol Learning, a decentralized approach to training and serving large models on consumer devices. We are a small, deeply technical team backed by Union Square Ventures, with a remote-first culture and team members spread globally.
Train models and evaluate AI systems to deliver world-class AI to the warfighter.
Run experiments in reinforcement learning, pretraining, multimodal training, and related areas.
Work as a self-starter in a fast-paced Series-A startup, interacting with customers and owning projects.
EdgeRunner AI builds secure, edge-deployed AI that helps American warfighters make faster decisions and interact with robotic platforms using natural language. It's a Series-A startup with a remote-first culture and a team of researchers and ML engineers.
Build the Learning Gym, a sandboxed environment for training and evaluating AI agents on real Niural workflows with ground-truth reward signals.
Engineer verifiers, run reinforcement learning loops, and establish honest baselines to prove agent improvements generalize.
Write first-author research papers and internal technical reports, and represent the work externally through preprints and talks.
Niural is a global payroll, Employer of Record (EOR), Agent of Record (AOR), and Contractor Management platform. It is a relentlessly ambitious team focused on building foundational internet infrastructure for the AI era.
Design, run, and analyse large-scale RL post-training experiments such as RLHF and RL from verifiable rewards.
Research and develop methods that improve out-of-distribution generalization, including reward design, task diversity, and regularization.
Build held-out evaluations to measure generalization and catch reward hacking, and turn successful research into production training recipes.
Cohere is a leading security-first enterprise AI company building cutting-edge foundation models and end-to-end products for businesses. We are a global team of researchers, engineers, and designers passionate about their craft, headquartered in Toronto with offices worldwide.
Build post-training and evaluation systems for multimodal agents operating browsers and desktops.
Turn real workflows into environments, rewards, and experiments that drive reliable agent improvement.
Work directly with customers and research partners while maintaining a substantive research agenda.
Prime Intellect is building the open superintelligence stack, providing infrastructure for post-training frontier AI models. The company has raised $150M in funding and fosters a culture of ownership and collaboration for ambitious AI teams.
Build and optimize RL infrastructure for large-scale training.
Improve efficiency across compute, memory, networking, and scheduling.
Design low-level optimizations, from kernels to runtime.
Prime Intellect builds the open superintelligence stack, providing full-stack infrastructure for frontier-scale AI post-training and deployment. The company has raised $150M from top investors and is mission-driven, focused on open-source AI and accelerating science.
Design and build challenging evaluations and executable environments for coding and long-horizon agents.
Develop scoring logic and verifiers to detect reward hacks and trace failures through model outputs and harness code.
Build and scale RL data pipelines for generating and curating challenging coding and agentic tasks.
Prime Intellect is building the open superintelligence stack, providing infrastructure for post-training frontier AI models and open-source models trained for long-horizon tasks. The company has raised $150M from top investors and is a mission-driven team passionate about accelerating science and AI.
Build, train, and evaluate machine learning models end to end, from data and experiment design through training and reporting.
Design reinforcement learning environments, reward functions, and verifiers, and fine-tune language models using methods like PPO, DPO, and GRPO.
Collaborate with domain experts, analyze datasets, and present findings clearly to both technical and non-technical audiences.
EnableComp provides specialty revenue cycle management solutions for healthcare organizations using its unified E360 RCM platform. It is a multi-year Top Workplaces recipient and has been named to the Inc. 5000 list of fastest-growing private companies for eleven years.
Build and post-train agents for scientific research, including literature search, code execution, and domain tool use.
Develop RL and multi-agent methods for proposing, testing, and critiquing hypotheses, with strong baselines and evaluation.
Design verifiers, reward signals, and reproducible experimental loops for scientific workflows, ensuring evidence-based improvements.
Prime Intellect builds the open superintelligence stack, providing infrastructure for frontier AI labs to train, evaluate, and deploy models at scale. The company has raised $150M from top investors and operators, and is building a culture focused on frontier research, real infrastructure, and go-to-market for a new category.
Develop AI methods and algorithms for reliable agents.
Design simulations and evaluations for agent behavior.
Partner with product and engineering to ship research.
The company builds an AI agent platform focused on making agents reliable in production. It is a research-driven environment with a collaborative, startup-oriented culture.
Build inference systems powering RL research, scaling toward 1–3T parameter models on NVL72 clusters.
Optimize rollout throughput, multi-node MoE inference, and weight-update overhead for long-horizon agents.
Collaborate with model researchers on architecture choices balancing model quality and execution cost.
Prime Intellect is building the open superintelligence stack, unifying compute, environments, and deployment for frontier-scale AI post-training. Backed by $150M in funding, the team is mission-driven and hard-working, focused on accelerating open AI research.
Design and execute experiments in reasoning model training, post-training alignment, test-time scaling, and systematic model evaluation for crypto-native applications.
Review and synthesize recent academic literature from NeurIPS, ICML, ICLR, and ACL to inform ongoing research directions in reasoning, alignment, and agentic AI.
Implement model variants and evaluation protocols using PyTorch and the Hugging Face ecosystem, tracking experiments with Weights & Biases.
Binance is a leading global blockchain ecosystem behind the world's largest cryptocurrency exchange by trading volume and registered users. Trusted by 300+ million people in 100+ countries, the company has a flat, results-driven culture and focuses on increasing the freedom of money and financial access worldwide.
Own applied RL programs from ambiguous customer problem to deployed improvement loop.
Understand workflows, build environments and evaluations, run post-training experiments, and deploy results.
Work directly with customer researchers, engineers, and leaders, accountable for research result and practical impact.
Prime Intellect builds the open superintelligence stack, providing a full-stack platform for frontier-scale post-training of AI models. The company has raised $150M from top investors and is seeking builders who want to work at the intersection of frontier research, real infrastructure, and go-to-market.
Develop and improve methods, algorithms, and systems across the lifecycle of reliable AI agents.
Create new approaches for simulating, evaluating, and optimizing agent behavior in real-world settings.
Partner with product and engineering teams to bring AI research into practical applications.
The company builds infrastructure for reliable AI agents in production, targeting enterprise workflows. It is a focused team of researchers and engineers working collaboratively on advancing agentic AI.
Build quantitative-research environments, agents, and post-training methods for systematic decision-making.
Turn workflows into tasks, run experiments against real baselines, and deploy what works.
Work directly with customers and quantitative funds while maintaining a substantive research agenda.
Prime Intellect builds the open superintelligence stack for post-training AI, unifying compute, environments, evaluations, and deployment. They have raised $150M from top investors and are building a culture of ambitious, externally-facing applied research.
Monitor and synthesize trends across AI and machine learning research, foundation models, agents, and data enrichment methods.
Tinker with emerging models, tools, and datasets to build prototypes and run experiments that turn ideas into evidence.
Identify high-leverage opportunities to enrich data assets and translate research into actionable recommendations.
Troveo builds the data platform AI labs need to train next-generation models, with the world's largest licensed platform of scarce, proprietary data. It is a small, high-impact team backed by top investors, focused on transparency and work-life balance.
Develop and improve AI methods, algorithms, and systems across the lifecycle of reliable agents.
Create new approaches for simulating, evaluating, and optimizing agent behavior in real-world settings.
Turn research ideas into working systems, prototypes, and production-facing capabilities.
The company is a research-focused team developing methods and systems to make AI agents more reliable in production. It operates as a small team with a culture of original research, experimentation, and close collaboration with product and engineering.