https://docs.edgelab.one/ Engineering paths, system maps, evaluation-bench writeups, operations notes, announcements, and reference pages about EdgeLab. en-us Tue, 26 May 2026 07:03:53 GMT https://docs.edgelab.one/guides/follow-a-portfolio-mutation/ https://docs.edgelab.one/guides/follow-a-portfolio-mutation/ Tue, 26 May 2026 21:55:00 GMT Portfolio healing starts with observation and diagnosis, then moves through research-cell admission, paper realization, shadow behavior, and promotion or rejection. https://docs.edgelab.one/guides/follow-an-eval/ https://docs.edgelab.one/guides/follow-an-eval/ Tue, 26 May 2026 21:50:00 GMT An EdgeLab eval starts as a dataset case, runs through a task function and agent/runtime, then becomes evaluator output that can gate research behavior. https://docs.edgelab.one/guides/follow-a-bar/ https://docs.edgelab.one/guides/follow-a-bar/ Tue, 26 May 2026 21:45:00 GMT A live bar moves from Alpaca into one-minute storage, RabbitMQ, ValKey aggregation state, indicator tables, and the execution queue before it can influence a strategy. https://docs.edgelab.one/guides/follow-a-research-idea/ https://docs.edgelab.one/guides/follow-a-research-idea/ Tue, 26 May 2026 21:20:00 GMT A research idea moves from scanner evidence to replay, verdicts, admission state, and paper realization before it can become portfolio behavior. https://docs.edgelab.one/system-map/runtime-map/ https://docs.edgelab.one/system-map/runtime-map/ Tue, 26 May 2026 21:15:00 GMT A map of the API, engine, queues, stores, broker paths, research loop, and public docs runtime as one connected system. https://docs.edgelab.one/evaluation-bench/inside-edgelab-evaluation-bench/ https://docs.edgelab.one/evaluation-bench/inside-edgelab-evaluation-bench/ Tue, 26 May 2026 21:00:00 GMT EdgeLab evaluates agents and strategies with datasets, task functions, custom evaluators, hard gates, and dogfood scoreboards before research can mutate portfolio state. https://docs.edgelab.one/engineering/from-bar-to-decision/ https://docs.edgelab.one/engineering/from-bar-to-decision/ Tue, 26 May 2026 20:10:00 GMT A concrete walkthrough of how EdgeLab processes a live market bar into indicators, regime context, strategy signals, risk checks, and position writes. https://docs.edgelab.one/engineering/how-edgelab-was-built/ https://docs.edgelab.one/engineering/how-edgelab-was-built/ Tue, 26 May 2026 20:00:00 GMT The actual platform shape behind EdgeLab: Go control plane, Python engine workers, market data ingestion, bar orchestration, execution, research cells, and the early portfolio-healing loop. https://docs.edgelab.one/systems/research-cells-and-portfolio-healing/ https://docs.edgelab.one/systems/research-cells-and-portfolio-healing/ Tue, 26 May 2026 19:50:00 GMT The emerging contract for turning research evidence into controlled paper observation and, eventually, portfolio mutation. What exists today and what is deliberately not claimed yet. https://docs.edgelab.one/systems/jobs-and-research-loops/ https://docs.edgelab.one/systems/jobs-and-research-loops/ Tue, 26 May 2026 19:40:00 GMT Scheduled jobs and queue consumers that keep EdgeLab evaluating strategies, taking snapshots, running CPO/ML optimization, and pushing research loops forward. https://docs.edgelab.one/systems/execution-runtime/ https://docs.edgelab.one/systems/execution-runtime/ Tue, 26 May 2026 19:30:00 GMT How indicator-enriched bars become strategy calls, risk checks, simulated fills, position locks, and paper/live position records. https://docs.edgelab.one/systems/market-data-bar-pipeline/ https://docs.edgelab.one/systems/market-data-bar-pipeline/ Tue, 26 May 2026 19:20:00 GMT How live equity and crypto bars move from Alpaca into TimescaleDB, RabbitMQ, ValKey aggregation state, indicator tables, and the execution queue. https://docs.edgelab.one/systems/engine-runtime/ https://docs.edgelab.one/systems/engine-runtime/ Tue, 26 May 2026 19:10:00 GMT The Python engine is a RabbitMQ-driven worker runtime for analysis, validation, execution, thesis workflows, trade analysis, knowledge-bank ingest, and AutoResearch allocation jobs. https://docs.edgelab.one/operations/docs-runtime/ https://docs.edgelab.one/operations/docs-runtime/ Tue, 26 May 2026 17:00:00 GMT The public docs site is a static Astro build served by nginx and deployed to the `docs` Kubernetes namespace. https://docs.edgelab.one/field-notes/docs-deploy-and-health/ https://docs.edgelab.one/field-notes/docs-deploy-and-health/ Tue, 26 May 2026 16:40:00 GMT The docs site is a static Astro build served by nginx. The Kubernetes probes and container health check both hit `/health`. https://docs.edgelab.one/announcements/engineering-publication-refactor/ https://docs.edgelab.one/announcements/engineering-publication-refactor/ Tue, 26 May 2026 16:20:00 GMT The public docs site is being split into engineering essays, evaluations, system references, field notes, and announcements. https://docs.edgelab.one/evaluations/how-edgelab-evaluates-strategies/ https://docs.edgelab.one/evaluations/how-edgelab-evaluates-strategies/ Tue, 26 May 2026 16:00:00 GMT Strategy evaluation needs baselines, gates, walk-forward checks, and explicit admission decisions. A single backtest score is not enough. https://docs.edgelab.one/engineering/knowledge-bank-retrieval-system/ https://docs.edgelab.one/engineering/knowledge-bank-retrieval-system/ Tue, 26 May 2026 15:40:00 GMT Knowledge Bank is a retrieval subsystem with intake, parsing, chunking, embeddings, and hybrid search. Its design is mostly about failure handling. https://docs.edgelab.one/engineering/autoresearch-keep-discard-evidence/ https://docs.edgelab.one/engineering/autoresearch-keep-discard-evidence/ Tue, 26 May 2026 15:20:00 GMT AutoResearch is a keep-or-discard loop for strategy changes. The useful part is not autonomy by itself; it is the evidence trail and stop rule.