Pinned Loading

1. alfred-infra alfred-infra Public

   AI-infrastructure hardening kit for multi-machine local-LLM clusters: monitoring (node_exporter, DCGM, Prometheus), Grafana dashboards, cold backups, network-binding audits

   Shell 1

2. alfred-rag alfred-rag Public

   Hybrid RAG stack: dense (LanceDB) + BM25 (Tantivy) + RRF + Qwen3-Reranker-8B. With sentence-transformers fine-tuning recipe for Qwen3-Embedding-8B.

   Python 1

3. context-bench context-bench Public

   Context-window benchmark for local LLM serving endpoints (MLX, vLLM, llama.cpp). Measures throughput and memory pressure across 4k-128k context sizes.

   Python 3

4. Agentic-Trader Agentic-Trader Public

   LLM-assisted options swing trading: the model proposes trades, you approve each entry, risk rails + an exit manager keep the book bounded. Sanitized public release.

   Python 1

5. gordon-gateway gordon-gateway Public

   Unified LLM gateway: routing + blockops tool-call translation + OTLP observability in one config-driven proxy (alfred/walter modes)

   Python

6. trading-eval-harness trading-eval-harness Public

   Eval harness for an options-trading LLM: 5-field scorer (move/call/vol/vs-IV/exp-move), directional accuracy, and conviction calibration buckets.

   Python 1 1