Articles
Notes on production AI, integrations, and platform engineering
Field notes on shipping production AI and RAG systems, enterprise integrations, distributed state, and the engineering judgment required to take software from prototype to production — plus what I've learned leading technical delivery.
Featured · Production AI Engineering
Why Most Enterprise AI Failures Are Integration Problems, Not Model Problems
Enterprise AI rarely dies at the model. It dies in the integration layer: identity, third-party APIs that disagree with their own docs, distributed state, and writing into systems of record. The discipline that ships integrations is the discipline that ships production AI.
- Enterprise AI
- Systems integration
- Production AI
- Distributed systems
- LLM applications
- Solution architecture
Production Voice AI · July 24, 2026 · 14 min read
What It Actually Takes to Deploy an AI Voice Agent Into a Live Call Center
A practical breakdown of deploying AI voice agents into live call operations — latency, interruptions, identity resolution, CRM integrations, booking workflows, observability, and human escalation.
→B2B SaaS Leadership · July 25, 2026 · 10 min read
The Technical Objection Is Often Not the Real Objection
How technical leaders distinguish real engineering objections from political resistance expressed in engineering language — with a pushback diagnostic for finding what a 'security concern' is actually protecting.
→Engineering Perspective · July 25, 2026 · 11 min read
The Engineer's Human Advantage: Communication in the Age of AI
AI is compressing the cost of producing code. What it doesn't compress: choosing the right problem, earning trust, explaining why work matters, and aligning people around action. Why communication is becoming a core engineering skill.
→Agentic Engineering · July 24, 2026 · 11 min read
When Multi-Agent Orchestration Actually Improves Software Delivery
A practical framework for multi-agent coding systems: when decomposition helps, what coordination actually costs, handoff contracts, an evaluation framework, and a decision matrix for choosing an architecture.
→Claude Code · July 24, 2026 · 13 min read
How I Use Claude Code to Take a Feature from Discovery to Production
A production Claude Code workflow from operational problem to deployed feature: repository inspection, plan mode, permission boundaries, branch isolation, continuous validation, diff review, and implementation reports.
→Claude Code · July 24, 2026 · 12 min read
How to Give Claude Code Autonomy Without Losing Control
A risk-tiered permission design for Claude Code: branch and filesystem isolation, command allowlists, permission modes, sandboxing, stop conditions, and which tasks deserve full autonomy versus supervision.
→AI Development Tooling · July 24, 2026 · 12 min read
Configuring OpenCode to Work Across Multiple LLM Providers
How to configure OpenCode across multiple LLM providers: global vs. project config, credential handling, provider blocks, model selection, local models via OpenAI-compatible endpoints, and troubleshooting.
→The systems behind the writing
All case studies →AI Voice Agents
AI Voice Agents for Home-Service Call Handling
606 calls handled · 83 booked end-to-end · $50K+ completed revenue traced
Conversational AI / RAG
RAG Copilot for CSR & Sales Teams
72% → 92% inbound close rate · Retrieval-enforced guardrails · Production feedback loop
Multi-Tenant Infrastructure
Multi-Tenant Deployment Infrastructure for 50+ Customer Apps
50+ apps · 1,000+ releases · 35% faster release velocity