Series
Collections grouped by topic, meant to be read in order. 18 posts · 115
Go deeper with series
LLM agents
Beyond chatbots: designing, evaluating, securing, and orchestrating autonomous agent systems.
Kubernetes troubleshooting
Symptom-by-symptom diagnosis of common Kubernetes failures.
LLM application architecture
Gateways, data pipelines, and prompt versioning across every layer of an LLM app.
RAG, end to end
Retrieval-augmented generation from first principles through production, evaluation, and security.
AI systems economics
Architecture patterns that keep LLM costs in check: model choice, latency, and ROI.
Enterprise AI architecture
Blueprints for taking PoCs to production — legacy integration, governance, and multi-cloud design.
Series
Series
Series — a multi-part deep dive.
Series
Series — a multi-part deep dive.
Series
Series — a multi-part deep dive.
Series
Series — a multi-part deep dive.
AI governance & MLSecOps
Technical governance for compliance, audit trails, abuse detection, and security automation.
AI data architecture
Data products, Data Mesh, and automated quality for AI-ready infrastructure.
Business framework for AI adoption
Turn adoption into outcomes: KPIs, change management, and process diagnosis.
Vector DB
HNSW indexes, hybrid search, and multi-tenant operations for large-scale AI search.
LLM prompt engineering
CoT, ToT, ReAct, A/B tests, and versioning — treating prompts as engineering.
Edge AI deployment
On-device inference and model compression: quantization, TFLite, and realtime streams.
Industrial AI architecture
AI-OT gateways, realtime anomaly detection, and edge-cloud hybrids for IIoT.
AI agent reliability
Detect failures, recover automatically, and hold SLAs for non-deterministic agents.