Poolside

Poolside's public-sector page attributes to Hunted Labs CTO and co-founder Hayden Smith a statement that Poolside models changed Hunted's internal secure-software development for customers. This names an actual reported partner workflow rather than mere platform availability; it does not identify end customers, contract volumes or measured productivity outcomes.

Poolside's May 28, 2026 technical-program account states that it moved from prototype to an ATO-approved production deployment in an air-gapped, classified environment in under nine months. The account describes STIG-hardened systems and deployment inside the customer's security boundary. This completed production milestone is distinct from the previously recorded Hunted Labs internal software-development testimonial.

Poolside has released open-weight agentic coding models and developer products, but public evidence of broad enterprise production deployment remains limited relative to leading coding platforms.

Poolside research and engineering spans trillion-scale pretraining data, synthetic data generation, post-training and RL, coding-agent environments, evaluation, distributed training, GPU-scale inference and frontier coding models, with thousands of GPUs available to applied-research teams.

Poolside trains proprietary foundation models for agentic coding from scratch with its own data, infrastructure and reinforcement learning, including Laguna models with long context.

Poolside trains proprietary foundation models for agentic coding from scratch with its own data, infrastructure and reinforcement learning, including Laguna models with long context.

Poolside announced a $500 million fundraise and reported scaling its training cluster to 10,000 GPUs. The financing amount is distinct from the existing general training-infrastructure evidence and does not establish profitability.

Poolside has built substantial model-training infrastructure and research capacity for frontier coding models and open-weight releases.

Poolside has built substantial model-training infrastructure and research capacity for frontier coding models and open-weight releases.

Poolside roles span pretraining/data research, synthetic data, post-training, reinforcement learning, evaluations, inference infrastructure, GPU scheduling, distributed systems, agent sandboxes, multimodality and full-stack product engineering, supporting broad portability across frontier-model and ML-systems employers.

Foundation-model training, coding agents, reinforcement learning and systems skills transfer exceptionally broadly across frontier labs and developer AI companies.

Foundation-model training, coding agents, reinforcement learning and systems skills transfer exceptionally broadly across frontier labs and developer AI companies.