// blog
Writing & Thinking
Deep dives on AEM architecture, Adobe Experience Platform, agentic AI systems, and the engineering patterns that underpin them.
8 posts
Context Engineering in Production: Beyond Agent Memory
The discipline above agent memory: what actually loads into the context window each turn - compaction, note-taking, JIT retrieval, and why two teams disagree on sub-agent isolation.
Inside Kimi K3's Architecture: KDA, Attention Residuals, and the 1M Context That Actually Runs
A technical breakdown of Kimi K3 - Kimi Delta Attention, the 3:1 hybrid attention stack, Stable Latent MoE across GPUs, attention residuals, and the 896-expert design that make a 1M-token context economical.
Durable Execution for Agentic Workflows: Why Your Agent Needs a Checkpoint, Not Just a Retry
Retries replay your agent from scratch and re-fire side effects. Checkpoints resume from the last completed step. Here is the real spectrum from retry to durable execution, and how to pick.
Agent Memory in 2026: Mem0 vs Zep vs Letta vs LangMem
Architecture deep-dive of Mem0, Zep/Graphiti, Letta, and LangMem - what each optimises for, when to pick which, and which benchmark claims to ignore.
AEM as an MCP Server: AI-Driven Operations for Enterprise Content
How I built a self-hosted MCP server for AEM operations and governance - 15 tools, dual-path AEMaaCS/6.5 support, query safety, async jobs, and a clear production path for enterprise teams.
GitNexus Deep Dive: The Knowledge Graph for AI Coding Agents
How GitNexus gives AI coding agents real codebase awareness - the indexing pipeline, Leiden clustering, hybrid BM25+RRF search, and when to use it.
Semantic Caching in Production: Beyond the Hype
How I built a 4-layer semantic cache for e-commerce search - two-zone thresholds, adaptive TTL, Kafka invalidation, and six production pitfalls nobody documents.
AEM 6.5 Clients: Are You Planning Your Next Move Before Support Phases Change?
AEM 6.5's last service pack is SP26, with support ending February 2027. Here's an honest breakdown of your two paths - LTS or AEMaaCS - and why the organisations that wait are the ones I see scrambling.