Deschide articolul original într-un tab nou Schimbarea acestei opțiuni se aplică imediat și reîncarcă pagina.
LECTURA
Înapoi la articole →

Stanley Yang — Writing

stanleycyang.com/feed.xml
Vizitează site-ul

Cel mai recent articol

Ultima verificare

Urmăritori 0 urmăritori
Articole: 110

  1. Benchmark Quantized Agent Models for Quality, Memory, and Speed

    (Stanley Yang — Writing)

    A reproducible protocol for comparing weight, activation, and KV-cache quantization without inventing savings or hiding quality regressions.

  2. Benchmark PostgreSQL 18 Async I/O for RAG Workloads

    (Stanley Yang — Writing)

    A reproducible protocol for testing PostgreSQL 18 AIO across vector search, filtered retrieval, ingestion, vacuum, and cache states.

  3. Backpressure for AI Agent Systems

    (Stanley Yang — Writing)

    Design admission control, bounded queues, concurrency limits, deadlines, and graceful degradation for agent workloads.

  4. Automatic Prefix Caching for Agent Workloads

    (Stanley Yang — Writing)

    Make repeated agent prompts cacheable in vLLM without crossing tenant boundaries, and measure whether cache hits actually improve latency.

  5. AI Agent Memory with Postgres and pgvector: Schema, Retrieval, and Forgetting

    (Stanley Yang — Writing)

    Design tenant-safe agent memory with provenance, hybrid search, retention, deletion, and pgvector indexing that does not mistake every transcript for knowledge.

  6. AI Agent Memory Poisoning: Threat Model, Detection, and Recovery

    (Stanley Yang — Writing)

    Protect persistent agent memory with source trust, write gates, quarantine, contradiction checks, provenance, and complete recovery procedures.

  7. AI Agent Evals in CI: A Release Gate for Prompts, Models, and Tools

    (Stanley Yang — Writing)

    Build a provider-neutral agent evaluation suite with deterministic checks, repeated trials, safety cases, uncertainty, and a CI policy that catches regressions.

  8. Agent Swarms: When Multiple AI Agents Help—and When They Make Things Worse

    (Stanley Yang — Writing)

    A practical decision framework for choosing a multi-agent architecture, plus a minimal coordinator that keeps concurrency, budgets, and failure handling explicit.

  9. Adopting OpenTelemetry GenAI Semantic Conventions Without Breaking Observability

    (Stanley Yang — Writing)

    A migration playbook for adopting the evolving OpenTelemetry GenAI span, metric, and event schema safely.

  10. Accessible Agent Interfaces for Keyboard, Screen Reader, and Voice Users

    (Stanley Yang — Writing)

    Design streaming conversations, tool progress, approvals, errors, and voice controls that preserve focus, user control, and understandable status.