Notre avis
Guide pour optimiser les performances des index vectoriels en production, notamment la latence, le rappel et l'utilisation mémoire.
Points forts
- Approche méthodique avec balayage de paramètres
- Couverture des stratégies de quantification
- Focus sur les métriques concrètes (latence, rappel, QPS)
Limites
- Nécessite des données de validation et des métriques de charge réelles
- Ne couvre pas la conception complète du système de recherche
Lorsque vous devez régler les paramètres HNSW ou choisir des stratégies de quantification pour équilibrer vitesse et rappel.
Si vous avez un petit jeu de données nécessitant une recherche exacte (utilisez un index plat).
Analyse de sécurité
SûrThe skill provides high-level guidance for tuning vector indexes without any executable commands, network access, or destructive actions. It is purely advisory and poses no execution risk.
Aucun point d'attention détecté
Exemples
I need to optimize my HNSW index for a 10M vector dataset. Guide me through a parameter sweep for ef_construction, M, and ef_search to achieve sub-10ms latency with 95% recall.Help me choose between scalar quantization and product quantization for my vector index. My memory budget is 2GB for 5M 768-dimensional vectors, and I need p95 latency under 5ms.name: vector-index-tuning description: "Optimize vector index performance for latency, recall, and memory. Use when tuning HNSW parameters, selecting quantization strategies, or scaling vector search infrastructure." risk: unknown source: community date_added: "2026-02-27"
Vector Index Tuning
Guide to optimizing vector indexes for production performance.
Use this skill when
- Tuning HNSW parameters
- Implementing quantization
- Optimizing memory usage
- Reducing search latency
- Balancing recall vs speed
- Scaling to billions of vectors
Do not use this skill when
- You only need exact search on small datasets (use a flat index)
- You lack workload metrics or ground truth to validate recall
- You need end-to-end retrieval system design beyond index tuning
Instructions
- Gather workload targets (latency, recall, QPS), data size, and memory budget.
- Choose an index type and establish a baseline with default parameters.
- Benchmark parameter sweeps using real queries and track recall, latency, and memory.
- Validate changes on a staging dataset before rolling out to production.
Refer to resources/implementation-playbook.md for detailed patterns, checklists, and templates.
Safety
- Avoid reindexing in production without a rollback plan.
- Validate changes under realistic load before applying globally.
- Track recall regressions and revert if quality drops.
Resources
resources/implementation-playbook.mdfor detailed patterns, checklists, and templates.
Ingénierie de Prompts
Data & IA
Bonnes pratiques et templates de prompt engineering pour maximiser les résultats IA.
Visualisation de Données
Data & IA
Génère des visualisations de données et graphiques adaptés à vos données.
Architecture RAG
Data & IA
Guide de configuration d'architectures RAG (Retrieval-Augmented Generation).