Writing
Every piece is a ten-minute read or less, with drawings: what broke and how it was fixed, one idea explained plainly, a decision with its trade-offs, and short briefs for the people who fund and run these systems.
- Source semantic understanding: a key to data program successExplainer · Data Engineering · Builders · 6 min read · 2026-11-06
- Shipping agents: versioning prompts, models and tools togetherExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- DBOS vs Temporal: durable agents without the dramaDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- Building agents on a moving ship: what has settled, what is still open, and how not to get stuckDecision · AI Engineering · Builders · 7 min read · 2026-10-31
- Rust or Python for agents? Probably bothDecision · AI Engineering · Builders · 7 min read · 2026-10-31
- Choosing your agent loop: LangGraph vs OpenAI Agents SDK vs Claude Agent SDK vs plain codeDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- Streaming on Databricks without the painDecision · Data Engineering · Builders · 8 min read · 2026-10-31
- CI/CD for Databricks with Asset BundlesExplainer · Data Engineering · Builders · 8 min read · 2026-10-31
- Agent-to-agent protocols (A2A): worth watching, not yet settledDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- Describe the world, not the testExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- What repeats is brokenExplainer · Data Engineering · Builders · 7 min read · 2026-10-31
- Why we wrote our own agent loopDecision · AI Engineering · Builders · 7 min read · 2026-10-31
- Your Databricks bill has a leak: 10 cost mistakes we keep seeingField note · Data Engineering · Builders · 8 min read · 2026-10-31
- Fail fast, own the result: critical thinking for forward-deployed engineersExplainer · Career · Builders · 8 min read · 2026-10-31
- The LLM gateway: the switch between your agent and every modelExplainer · LLM Engineering · Builders · 8 min read · 2026-10-31
- Measuring agent ROI: what to count before you scaleLeader brief · AI Engineering · Leaders · 6 min read · 2026-10-31
- From data engineer to AI engineer: the 6-month switchExplainer · Career · Builders · 7 min read · 2026-10-31
- Unity Catalog done right: one governance model for data and AIExplainer · Data Engineering · Builders · 8 min read · 2026-10-31
- Bronze, Silver, Gold, and where most teams get it wrongExplainer · Data Engineering · Builders · 4 min read · 2026-10-31
- Your agent has the keys: security architecture for autonomous agentsExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- The autonomous agent tech stack, one concern at a timeExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- Just use Postgres: the one database your agent needsDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- The horse and the harnessLeader brief · AI Engineering · Leaders · 5 min read · 2026-10-31
- Every guardrail your agent needs, sorted: must-have, should-have, rule or judgeExplainer · AI Engineering · Builders · 5 min read · 2026-10-31
- AgentTwin vs LangWatch, Veris and friends: the simulation landscapeDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- Before it ships: offline evals for autonomous agentsExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- What one conversation costs: under a tenth of a centLeader brief · Cost · Leaders · 5 min read · 2026-10-31
- Should you build or buy your agent? A one-page guide for leadersLeader brief · AI Engineering · Leaders · 6 min read · 2026-10-31
- Prompt, RAG or fine-tune? Pick in that orderDecision · LLM Engineering · Builders · 7 min read · 2026-10-31
- Swap anything, break nothing: software design patterns for agent harnessesExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- A flight simulator for your agent: what AgentTwin is, and what it is notExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- The AI engineer roadmap: what to learn, in what order, and what to skipExplainer · Career · Builders · 8 min read · 2026-10-31
- Who checks the checker? The oracle problem in agent simulationExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- How to learn in the AI era: draw the map, grow a T, and teach it backExplainer · Career · Builders · 7 min read · 2026-10-31
- Talk to your data the right way: Genie and metric views on DatabricksExplainer · Data + AI · Builders · 7 min read · 2026-10-31
- The agent testing pyramidExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- Data contracts: stop breaking the team downstreamExplainer · Data Engineering · Builders · 7 min read · 2026-10-31
- What a forward-deployed engineer actually doesExplainer · Career · Builders · 9 min read · 2026-10-31
- There is no best model, only the best model for this stepDecision · LLM Engineering · Builders · 8 min read · 2026-10-31
- Your Databricks cluster scales up and never comes back downField note · Data Engineering · Builders · 8 min read · 2026-10-31
- Your agent passed every test. Then it met a real customer.Field note · AI Engineering · Builders · 7 min read · 2026-10-31
- Delta, Iceberg, or both? Open table formats are settledDecision · Data Engineering · Builders · 8 min read · 2026-10-31
- Stop writing pipelines: one metadata-driven ingestion framework for 500 tablesExplainer · Data Engineering · Builders · 7 min read · 2026-10-31
- Under the hood: why fundamentals are the slow road that winsExplainer · Career · Builders · 6 min read · 2026-10-31
- Data quality as config, not codeExplainer · Data Engineering · Builders · 7 min read · 2026-10-31
- The model is 10%. The harness is the other 90%.Explainer · AI Engineering · Builders · 6 min read · 2026-10-31
- One agent or many? When multi-agent is a mistakeDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- Most of an agent is ordinary softwareBuild log · AI Engineering · Builders · 7 min read · 2026-10-31
- Human in the loop without humans in the wayExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- Answer without the model first: 104 ms vs 7 secondsBuild log · AI Engineering · Builders · 7 min read · 2026-10-31
- Seatbelts for agents: guardrails that actually holdExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- It passed offline. Now watch it live: online evals and monitoringExplainer · AI Engineering · Builders · 9 min read · 2026-10-31
- One harness, five homes: the agent harness on Azure, GCP, AWS, Databricks and open sourceDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- OpenTelemetry is here to stay: instrument your agent once, send traces anywhereDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- Why your agent costs $4 per question: 9 cost mistakesField note · AI Engineering · Builders · 8 min read · 2026-10-31
- Build only the gaps: AgentTwin assembled from open-source toolsDecision · AI Engineering · Builders · 7 min read · 2026-10-31
- If you can't replay it, you can't fix it: tracing autonomous agentsExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- Agent memory without the magicExplainer · AI Engineering · Builders · 8 min read · 2026-10-31
- MCP is here to stay: give your agent tools the standard wayDecision · AI Engineering · Builders · 8 min read · 2026-10-31
- Context engineering: what goes into the window, and what stays outExplainer · AI Engineering · Builders · 8 min read · 2026-10-31