Writing

32 essays · 2018–2026

On the decisions that shape enterprise AI, data platforms, and the organizations that run them.

Latest · AI Strategy AI Did Not Make Engineering Cheaper. It Made It Bigger. Most CFOs read AI as a reason to shrink engineering. The work that got cheaper was first-draft code. The work that makes software safe to ship became the new bottleneck. Treat AI as freed capacity, not a cost-out lever. June 2026 · 11 min readRead →
O'Reilly Radar Who Authorized That? The Delegation Problem in Multi-Agent AI When an orchestrator delegates to a specialist and the specialist calls a tool, authorization quietly widens at every hop. Privilege concentrates at the agent that touches the outside world, and the prompt is not an authorization model. May 2026 · 11 min read · oreilly.com↗ Multi-Agent AI The Model Is Not the Product. The Harness Is. Anthropic gave the industry the language for the layer it has been buying without a label. The harness is the platform. The model is the part you stopped paying for. May 2026 · 10 min read→ Multi-Agent AI The Agent Demo Gap Multi-agent demos go viral. Multi-agent production gets a six-month security review. Both reactions are correct. The strategic move is to ship the demo loop with the production primitives baked in, not bolted on. May 2026 · 7 min read→ Multi-Agent AI The Same Database, Twice Two AI agents, two production database deletions, nine months apart, in the same costume. The trust model that failed both times, and the layer that should have been there. May 2026 · 12 min read→ Multi-Agent AI Add Cryptographic Identity to Your AI Agents in 5 Minutes Your agents call tools anonymously. The Agent Identity Protocol gives each agent an Ed25519 keypair, scoped delegation tokens, and audit-ready chains. Works with Google ADK, CrewAI, and LangChain. April 2026 · 6 min read→ Multi-Agent AI Building a Policy Gateway for AI Agent Delegation AIP Gateway is a drop-in proxy that verifies who delegated what before a tool executes. YAML policy, audit-grade logs, no agent rewrites. Includes a loan origination demo showing scope, chain, and maker-checker enforcement. 2026 · 4 min read→ Multi-Agent AI The Agent Identity Gap: Why MCP and A2A Need Verifiable Delegation A scan of ~2,000 MCP servers found all lacked authentication. OAuth 2.1 covers single-hop auth, but multi-agent delegation chains need more. How verifiable identity connects to routing, provenance, and reasoning. 2026 · 4 min read→ Multi-Agent AI When Smarter AI Isn't Worth It Most teams over-engineer simple tasks and under-engineer complex ones. A framework for matching AI architecture complexity to task complexity, with data on when 62x more compute is the right call. 2026 · 18 min read→ Multi-Agent AI The Missing Agent Stack: Identity + Durable Execution with LDP and JamJet Code-first walkthrough: build identity-aware agent routing with LDP identity cards and JamJet's Coordinator. Difficulty-based scoring, provenance-weighted synthesis, runnable example. 2026 · 6 min read→ Multi-Agent AI From Debate to Deliberation: When Multi-Agent Reasoning Needs Structure Multi-agent debate is unstructured and lossy. DCI introduces typed reasoning moves, preserved disagreements, and guaranteed convergence, with honest results on when the 62x token cost is justified. 2026 · 7 min read→ Multi-Agent AI Why Multi-Agent AI Systems Need Identity-Aware Routing Current agent protocols treat all models as interchangeable black boxes. LDP introduces identity-aware delegation, and the results show where it matters and where it doesn't. 2026 · 10 min read→ Multi-Agent AI From NER Pipelines to LLM Agents: How Production NLP Changed in Seven Years Seven years from BiLSTM-CRF at a startup to multi-agent protocols: what changed in production NLP, what stayed the same, and what the arc tells us about where we are going. March 2025 · 7 min read→ Enterprise AI & Governance Why Agent Strategy Becomes an Architecture Problem at Scale Enterprises are repeating the microservices mistake with agents: letting teams adopt without defining the control plane. Five cross-cutting concerns only architecture can govern, and three artifacts every architecture function must create. April 2026 · 6 min read→ Enterprise AI & Governance Your ML Risk Framework Wasn't Built for GenAI. Here's What's Missing. Why traditional model risk management fails for LLMs: hallucination policy, GenAI risk dimensions, deployment gates, prompt audit trails, and what a complete governance framework looks like. 2026 · 12 min read→ Enterprise AI & Governance The Year LLMs Met Compliance — And Compliance Wasn't Ready GPT-4 is genuinely capable. Enterprise wants to use it. But the governance frameworks built for classical ML, model validation, risk management, audit trails, were not designed for non-deterministic models with emergent capabilities. October 2023 · 11 min read→ Enterprise AI & Governance GPT-3 Changed the Game — Is Enterprise Ready? 175 billion parameters, few-shot learning, and an API. GPT-3 shows where language models are going. But cost, latency, data privacy, and governance stand between the demo and production. November 2020 · 7 min read→ AI Leadership & Strategy Most Companies Know Their AI Spend. Almost None Know Their AI Readiness. Most organizations know their AI spend. Almost none can answer basic questions about whether they're ready to absorb what they're buying. A diagnostic across five readiness dimensions, and a 10-minute assessment tool to find the gaps. 2026 · 5 min read→ AI Leadership & Strategy The Middle Management AI Gap Boards are excited about AI. Engineers are building it. But the layer between the boardroom and the engineering floor is where most enterprise AI programs quietly die, and nobody talks about it. 2026 · 8 min read→ AI Leadership & Strategy Why Most Enterprise AI Programs Fail Before They Start The failure mode isn't technical. It's organizational. Most enterprises skip the operating model and jump straight to tooling, then wonder why nothing scales past the pilot. 2026 · 4 min read→ AI Leadership & Strategy The AI Center of Excellence Is Dead. Long Live the AI Operating Model. Centers of Excellence centralize capability. Operating models distribute it. The difference determines whether AI scales or stalls at the enterprise level. November 2025 · 4 min read→ AI Leadership & Strategy Why I Chose Regulated AI Over Startup Speed After three years building AI at a startup, I moved to a global bank. Not despite the constraints, because of them. Regulated environments are where AI gets hardest and most valuable. August 2021 · 6 min read→ Data Architecture Why We Chose dbt Over BigQuery Dataform JavaScript supply chain risk, SQL-first philosophy, multi-warehouse portability, and what actually matters when choosing a transformation framework in a regulated enterprise. 2026 · 11 min read→ Data Architecture Data Vault 2.0 in Banking: Architecture for the Audit That Hasn't Happened Yet Why insert-only modeling, hash keys, and full lineage aren't academic exercises, they're survival patterns in regulated environments where every record must be defensible. 2026 · 4 min read→ Data Architecture Building Data Foundations While Everyone Chased Models The industry is obsessed with models. Meanwhile, the actual bottleneck for enterprise AI is data infrastructure: Data Vault 2.0, BCBS 239, lineage, and the unsexy work that makes everything else possible. March 2022 · 9 min read→ NLP & Machine Learning RAG in 2026: From Pipeline to Agent Two years on, the retrieve-then-generate pipeline became a loop the model drives. What agentic retrieval changed for long context, evaluation, and guardrails, plus a production checklist for building it now. June 2026 · 13 min read→ NLP & Machine Learning We Switched to PyTorch in 2020. Was It the Right Call? Four years after switching from TensorFlow to PyTorch during the pandemic: what happened to TF, why PyTorch won, and why inference engines like vLLM matter more than training frameworks now. October 2024 · 6 min read→ NLP & Machine Learning RAG in Production: What Breaks When You Move Past the Tutorial Every RAG tutorial works. Then you move to production and discover that chunking strategy matters more than embedding models, retrieval quality matters more than generation, and evaluation is the hardest problem. June 2024 · 10 min read→ NLP & Machine Learning Switching from TensorFlow to PyTorch — A Practical Assessment We switched our NLP projects from TensorFlow 1.x to PyTorch mid-pandemic. Eager execution, HuggingFace, debugging, and what we gave up. A practical framework comparison for a small team. June 2020 · 8 min read→ NLP & Machine Learning Building NLP Pipelines Before Transformers Were Easy What production NER actually looks like: BiLSTM-CRF, character embeddings, gazetteers, and the pipeline engineering that matters more than the model. Lessons from building NLP systems at a startup. June 2019 · 9 min read→ NLP & Machine Learning Attention Is All You Need — A Practitioner's Guide to the Transformer Breaking down the paper that replaced RNNs, LSTMs, and sequence-to-sequence models with a single mechanism: attention. What the Transformer does, why it works, and why it matters. September 2018 · 9 min read→ NLP & Machine Learning Classifying 7,000 Product Codes from Four Words of Text Comparing LSTM, BiLSTM, GRU, CNN, and hybrid architectures for HS code classification from short product descriptions. What worked, what didn't, and why CNNs beat RNNs on short text. August 2018 · 11 min read→