Open source developer and startup founder

github

Luke Hinds  • Peerlist

DeepFabric is a specialised dataset generation and model fine-tuning framework designed for training small language models (SLMs) to become highly efficient and capable Agents. By combining reasoning traces with tool calling patterns, and enforcement of type based structured outputs - DeepFabric enables you to fine-tune models that make intelligent decisions, select appropriate tools, and execute multi-step workflows efficiently and with accuracy.Built for ML engineers, researchers, and AI developers, DeepFabric streamlines the entire agent training pipeline: from hierarchical topic generation to structured reasoning templates to model-ready formats across all major training frameworks. Whether you're building MCP-compatible agents, distilling capabilities into smaller models, or creating specialized tool-calling systems, DeepFabric provides the high-quality, diverse training data you need at scale.