{"slug": "end-to-end-infrastructure-for-training-and-inferencing-open-weight-models", "title": "End-to-end infrastructure for training and inferencing open weight models", "summary": "AC2, an end-to-end infrastructure platform for training and inferencing open weight models, has launched with SDK, runtime, and tracing packages that support coding agents like Claude Code, Cursor, and Codex. The platform enables users to evaluate agents, train models using GRPO, supervised fine-tuning, or on-policy self-distillation, and route inference through policies and endpoints.", "body_md": "## Onboard with your coding agent\n\nThe fastest way to get started is to let your coding agent set AC2 up for you. Paste this prompt into Claude Code, Cursor, Codex, or any coding agent:\n[Quickstart](/quickstart).\n\n## Packages\n\n- **`ac2.sdk`** — platform client for datasets, workloads, sessions, traces, and secrets.\n- **`ac2.runtime`** — agent framework for agents, environments, orchestrators, tasks, graders, and episodes.\n- **`ac2.tracing`** — OpenTelemetry-based observability SDK. Add`@traced` to any function to capture timing, inputs, outputs, and nested spans. Exports to ClickHouse and Langfuse.\n\n## Quickstart\n\nInstall the SDK and train a model.\n\n## Runtime\n\nDefine agents, environments, orchestrators, and eval tasks.\n\n## SDK\n\nManage datasets, deployments, runs, traces, and secrets from Python.\n\n## MCP server\n\nAnalyze eval and train traces from your editor.\n\n## Choose a workflow\n\n## Evaluate an agent\n\nRun evals, grade traces, and compare behavior.\n\n## Train a model\n\nChoose GRPO, supervised fine-tuning, or on-policy self-distillation.\n\n## Route inference\n\nDeploy a model and route traffic through Policies and Endpoints.\n\n## FAQ\n\nFind answers for failed runs, datasets, package installation, and cleanup.", "url": "https://wpnews.pro/news/end-to-end-infrastructure-for-training-and-inferencing-open-weight-models", "canonical_source": "https://docs.appliedcompute.com", "published_at": "2026-09-04 21:25:49+00:00", "updated_at": "2026-09-08 11:03:05.809547+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-tools", "mlops", "ai-research"], "entities": ["AC2", "Claude Code", "Cursor", "Codex", "OpenTelemetry", "ClickHouse", "Langfuse"], "alternates": {"html": "https://wpnews.pro/news/end-to-end-infrastructure-for-training-and-inferencing-open-weight-models", "markdown": "https://wpnews.pro/news/end-to-end-infrastructure-for-training-and-inferencing-open-weight-models.md", "text": "https://wpnews.pro/news/end-to-end-infrastructure-for-training-and-inferencing-open-weight-models.txt", "jsonld": "https://wpnews.pro/news/end-to-end-infrastructure-for-training-and-inferencing-open-weight-models.jsonld"}}