Sam Griffith

Sam Griffith

Senior Software Engineer / Tech Lead

Building Cloud-Native Platforms - Leaning into Agentic AI

  • About
  • Experience
  • Projects
  • Contact
  • Articles
Daily digest

Articles

A daily briefing on AI, agentic systems, and software engineering — curated and synthesized by an autonomous agent, published here each morning. Articles expire after 90 days.

Today · 29 August 2026 Latest

Cutting the Fat: Cost and Latency Optimization for Agentic Pipelines in Production

How do you keep a multi-LLM, knowledge-graph-backed agentic pipeline fast and cheap when it's sitting in the hot path of a real-time customer-facing application?

Read today's digest
48 previous articles
  • Aug 27Cutting the Latency Between Agent Failure and Human Eyes: HITL Escalation Patterns for Production Support Systems
  • Aug 25Alerting Without Crying Wolf: SRE Strategies for Detecting Agent Drift Before It Becomes an Incident
  • Aug 23Critic Agents That Actually Catch Drift: Designing Feedback Loops for High-Stakes Agentic Systems in the First Six Weeks
  • Aug 21Model Versioning for Agentic Systems: Rolling Back Without Breaking What's Downstream
  • Aug 19Hardening Agent Tool Calls: Prompt Injection Defense and Supply Chain Trust in Multi-Tenant Agentic Systems
  • Aug 17Least-Privilege Tool Credentials in Agentic Systems: Designing Function Calling Protocols That Actually Hold Up in Production
  • Aug 15Holding the Line: How Senior Engineers Keep Architectural Judgment While AI Tooling Accelerates Everything Around Them
  • Aug 13Wiring Agents Into Event-Driven Systems Without Burning Down Your Consistency Guarantees
  • Aug 11Locking Down the Agent Layer: Rate Limiting, Input Validation, and Governance Patterns for Multi-Tenant LLM Platforms
  • Aug 9Context Window as a First-Class Resource: Resizing Strategies for AI Coding Agents Across Multi-File Sessions
  • Aug 7What Early Enterprise Adopters Actually Traded Away to Ship Agentic Systems
  • Aug 5Observing the Retrieval Orchestrator: Platform-Grade Monitoring for Agent-Based RAG Under Real Load
  • Aug 3Edge-First Agentic Systems: Caching Discipline, Delivery Guarantees, and the Orchestration Shift That Changes How You Design Both
  • Aug 1Model Routing Under Pressure: Designing Latency-Aware Fallback Strategies for Production Agentic Pipelines
  • Jul 31Contract-Driven Agents: How Platform Teams Can Enforce LLM Output Schemas Without Strangling Agent Autonomy
  • Jul 29Checkpointing Agent State at Failure Boundaries: Patterns for Long-Running Tasks That Don't Start Over
  • Jul 27Cache or Recompute? The Agentic Pipeline Decision That Determines Whether Your Real-Time Agent Feels Fast or Broken
  • Jul 25Handoff Without the Drop: Engineering Low-Latency Human Escalation for Agent-Based Customer Support
  • Jul 23Supervisor Topologies for Cascading Failure Detection: Building Self-Healing Agentic Pipelines Without Manual Intervention
  • Jul 21Alerting for Things You Can't Directly Measure: Detecting Hallucination and Pipeline Drift in Production Agentic Systems
  • Jul 19How to Test Agents That Will Lie to You, Leak Your Data, and Fail in Production
  • Jul 17Credential-Safe Tool Integration: Function Calling Protocols for Platform Teams Building Agentic Systems
  • Jul 15Decomposed Observability for LLM Pipelines: Catching Cascading Failures Before They Compound
  • Jul 13Phase-Emitting Agent Loops: Building Observability Frameworks That Catch Failures Before They Become Incidents
  • Jul 11Async Guardrails at Scale: How to Filter Agent Outputs Without Choking the Pipeline
  • Jul 9Wiring Agents Into Event Streams Without Losing the Wheel: Control Plane Patterns for Agentic Event-Driven Systems
  • Jul 7Building Contextual Anomaly Detection Into Customer-Facing Agents: A Security Architecture for Prompt Injection Defense
  • Jul 5Per-Phase Provenance Logging: Designing Audit Schemas for Compliant Financial Agent Pipelines
  • Jul 3MCP Scope Enforcement in VPC-Integrated Agentic Systems: Drawing Hard Lines Before Agents Cross Them
  • Jul 1RAG in Production: Typed Inputs, Elastic Storage, and the Latency Tradeoffs That Actually Matter
  • Jun 29Adaptive Agentic Integration: How to Wire Agents Into Microservices Without Painting Yourself Into a Corner
  • Jun 27When to Override the AI: A Platform Engineer's Guide to Maintaining Architectural Judgment Under AI Acceleration
  • Jun 25Cut the Fat, Keep the SLO: Latency and Cost Optimization Patterns for Large-Scale Agentic Recommendation Pipelines
  • Jun 23Contract-Driven Agents: The Output Schema Patterns That Make Agentic Systems Actually Trustworthy
  • Jun 21Contract-First Agent Design: How Typed Outputs Let You Ship New Agent Capabilities Without Breaking Production
  • Jun 19Agent Contracts in Production: How Structured Outputs Keep Multi-Agent Pipelines From Eating Themselves
  • Jun 17Contract-Driven Agents: How to Keep Human and AI Workflows Consistent When They Share the Same Data
  • Jun 16Checkpointing Agent Workflows: How to Stop Treating Interruption as Catastrophe
  • Jun 15Checkpoint Without Choking: Designing Agent State Persistence That Doesn't Kill Real-Time Customer Service
  • Jun 14Checkpoint What Matters: Designing Resumable Agent Workflows Without Drowning in State
  • Jun 12Designing Recoverable Agent Workflows: Step-Level Checkpoints and Localized State Boundaries
  • Jun 10Idempotent Agent Loops: Designing Fault-Tolerant Checkpoint Workflows
  • Jun 8Architecting Auditable State Checkpoints in Long-Running Agent Workflows
  • Jun 7Architecting Resilient Agent State: Checkpoint Recovery and Intent Governance in High-Stakes Pipelines
  • Jun 6State Without Memory Is Just Storage: Building Agent Checkpoints That Actually Tell You What Went Wrong
  • Jun 4Before You Ship: Designing Agent Regression Suites That Catch What Unit Tests Miss
  • Jun 3Hard Walls, Not Soft Nudges: Engineering Governance Boundaries That Hold in Production
  • Jun 2Trace the Decision, Not the Prompt: Building Compliance-Safe Agent Observability

Articles are generated automatically and expire after 90 days.

Open full page