Dan Austin
Specialising in agentic AI systems, reinforcement learning, and long-horizon tasks.
At a Glance
- AI Researchers
- Enterprise Engineering Teams
- Open Source Community
- Legal and Government Sectors (Consultancy)
AI Tools by Dan Austin
(1)
AI Trains AI
Nested RL Training Research Project
Discussions
No discussions yet
Be the first to start a discussion about Dan Austin
Latest News
Products & Services
An RL-trained agent system that automates the creation of RL training jobs (environments, rewards, datasets) and manages GPU training runs.
An RL-trained orchestrator model (based on Qwen3-14B) designed to coordinate explorer and coder subagents.
A novel multi-agent architecture for complex problem solving, outperforming Claude Code on TerminalBench.
GRPO-based RL training infrastructure for scaling coding agent training across multiple nodes.
Market Position
Highly specialized in agentic RL, with independent projects outperforming institutional benchmarks like Stanford's TerminalBench (ranking #12).
Leadership
Founders
Dan Austin
Principal Software Engineer at Microsoft; formerly Data Scientist at Fabric (Aviva Insurance). Specialized in agentic AI systems and reinforcement learning.
Executive Team
Dan Austin
Principal Software Engineer / Creator
Expert in Reinforcement Learning and agentic systems with experience at Microsoft and Aviva.
Founding Story
A career dedicated to pushing the boundaries of agentic AI and reinforcement learning, transitioning from corporate data science to leading engineering roles and independent research projects.
Target Markets
- AI Researchers
- Enterprise Engineering Teams
- Open Source Community
- Legal and Government Sectors (Consultancy)
- Automated RL training orchestration
- Complex multi-step coding problems
- Agentic data generation and validation
- Enterprise-scale AI agent deployment
- Microsoft
- Aviva
- Government of Alaska
- Flow Legal