About
CV
Projects
Sports
Blog

Blog – Insights on AI & ML

Articles and insights from Nolan Cacheux on Data Science, Machine Learning, and AI Engineering

Start here

The blog is organized around the work I actually ship: retrieval quality, agent reliability, and production ML operations.
RAG & evaluationRetrieval design, chunking, offline evals, and answer quality gates.
Agents & governanceEnterprise assistants, guardrails, escalation paths, and operating rules.
MLOps & observabilityDeployment, monitoring, cost, latency, traces, and incident response.
Featured article

RAG Evaluation Framework: Metrics That Predict Production Quality

April 9, 20267 min readRAG Evaluation
A practical RAG evaluation framework covering retrieval precision, grounded answer quality, citation correctness, and release gates.
Read featured article →

More articles

Evaluation Regression Alerting Blueprint for AI Products

April 13, 20268 min readAI Evaluation
How to detect quality drift early with evaluation regression alerts tied to release gates.
Read article →

LLM Cost Optimization with Quality Guardrails

April 12, 20268 min readLLM Cost
Reduce LLM spend with routing, caching, and prompt compression while preserving answer quality and user trust.
Read article →

Prompt Security and Tool Hardening Checklist

April 12, 20268 min readAI Security
A hardening checklist to reduce prompt injection, tool abuse, and unsafe output propagation in agentic AI systems.
Read article →

Multimodal RAG Production Playbook: Documents, Images, and Tables

April 11, 20268 min readMultimodal RAG
A production playbook for multimodal RAG systems handling PDFs, screenshots, and structured tables with grounded answers.
Read article →

Synthetic Data Pipeline for Domain Fine-Tuning

April 11, 20268 min readSynthetic Data
Design a synthetic data pipeline that improves model quality without poisoning production behavior or violating governance constraints.
Read article →

Agent Evaluation Flywheel: From Prototype to Reliable Production

April 10, 20268 min readAgent Evaluation
How to operationalize an evaluation flywheel for AI agents with release gates, regression suites, and business-aligned quality signals.
Read article →

vLLM Serving Blueprint: Low-Latency Inference at Scale

April 10, 20268 min readInference Engineering
A practical serving blueprint for vLLM in production: routing, KV cache strategy, concurrency limits, and latency SLO management.
Read article →

LLM Observability: Traces, Costs, and Quality Signals for Production

April 8, 20262 min readLLM Observability
A production LLM observability stack for tracing prompts, measuring response quality, controlling token spend, and reducing incident time.
Read article →

Agent Architecture Patterns for Reliable Enterprise AI Systems

April 7, 20262 min readAgent Architecture
A decision framework for agent architecture: when to use router-worker, planner-executor, graph orchestration, and deterministic guardrails.
Read article →

Enterprise AI Governance Framework: Controls Without Delivery Bottlenecks

April 6, 20262 min readAI Governance
A practical enterprise AI governance framework for policy enforcement, risk scoring, model lifecycle controls, and auditable releases.
Read article →

How I Build Production-Ready RAG Systems

March 28, 20262 min readRAG
The practical stack I use for enterprise RAG: retrieval quality, observability, eval loops, and guardrails.
Read article →

MLOps Checklist for Real Deployments

March 20, 20262 min readMLOps
A compact checklist to ship ML systems safely: data contracts, CI/CD, model registry, drift alerts, and rollback strategy.
Read article →

What Enterprise AI Teams Should Actually Measure

March 10, 20263 min readAI Strategy
Beyond vanity metrics: the KPIs that prove AI delivers business value in enterprise environments.
Read article →