Haitao Wang
VP of Engineering at Tacnode
Haitao Wang is VP of Engineering at Tacnode, where he leads the team building the Tacnode Context Lake™. He spent nearly two decades at Microsoft, eight years of it on the SQL Server engine, working on T-SQL programmability, the plan cache, engine performance and columnstore, and later managed the big data and shared data platform teams behind Bing Ads and Skype for Business. He writes about streaming ingestion, incremental materialized views, and what it takes to keep derived state current for decisions made in production.
Posts by Haitao (12)
![Incremental Materialized View: How to Keep Derived State Fresh in Real Time](data:image/svg+xml,%3Csvg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 1 1"%3E%3C/svg%3E)
Incremental Materialized View: How to Keep Derived State Fresh in Real Time
Standard materialized views recompute everything on a schedule. Incremental materialized views apply only the delta — bounding refresh cost to what actually changed. Here's how they work across Postgres, ClickHouse, Databricks, streaming databases, and Tacnode, what each approach can and can't do, and where they all hit the same structural limit.
Haitao Wang|Mar 24, 2026ETL Pipelines: What They Are, How They Work, and When to Eliminate Them
ETL pipelines extract data from source systems, transform it into a usable format, and load it into a destination. This guide covers how ETL pipelines work, common architectures, tools, failure modes, and when streaming and CDC approaches eliminate the need for batch ETL entirely.
Haitao Wang|Mar 13, 2026Streaming Database: What It Is, How It Works, and When You Need One
A streaming database replaces the batch query model with continuous computation — materialized views maintained incrementally as data arrives. Here's how they work, when they help, and where they run out of runway.
Haitao Wang|Mar 5, 2026LLM Model Staleness: Why Models Go Stale After Training
Your LLM answered correctly last month—now it's confidently wrong. This isn't hallucination. It's model staleness: a structural limitation that fine-tuning and RAG can't solve. Here's the architectural fix that actually works.
Haitao Wang|Jan 9, 2026Postgres Materialized Views: Create, Refresh, and Optimize
How to create, refresh, and optimize a Postgres materialized view — plus the structural limit every team hits when reads need fresh derived state under concurrency.
Haitao Wang|Apr 29, 2026Postgres CDC: Logical Replication, Debezium, and Change Data Capture Options
Postgres CDC explained: how change data capture works in PostgreSQL, with a runnable example, trigger and polling approaches, logical decoding with replication slots and publications, how to enable it on self-hosted, RDS, Aurora, Cloud SQL, Azure and Supabase, the replication slot risks that take databases down, and how native replication, Debezium, Flink CDC and managed services compare.
Haitao Wang|Sep 30, 2026Agent Double-Spend: Two Agents, One Budget
Agent double-spend: two AI agents check one budget, both see room, both spend. Why provider budget alerts miss it, and how reserve-then-commit stops it.
Haitao Wang|Oct 19, 2026Agentic TOCTOU: When an Agent Checks, Then Acts on Changed State
Agentic TOCTOU: an AI agent checks state, reasons for seconds, then acts after it changed. The research behind the term, the shared-state case, and the fixes.
Haitao Wang|Oct 19, 2026AI Agents and Schema Migrations: Two Agents, One Schema
When AI agents run schema migrations against one live database, both changes apply and the result can be wrong. Where it happens and how to coordinate.
Haitao Wang|Oct 19, 2026Durable Execution for AI Agents: Recovery Isn’t Coordination
What durable execution engines like Temporal, Restate, DBOS and Inngest guarantee for AI agents, and what they leave to you when many workflows share state.
Haitao Wang|Oct 19, 2026Multi-Agent Concurrency Control: Six Mechanisms Compared
Multi-agent concurrency control compared: locks, version checks, conditional writes, serializable transactions, queues and leases, and where each breaks.
Haitao Wang|Oct 19, 2026Race Conditions Between AI Agents
Race conditions between AI agents corrupt shared state. The four shapes the bug takes, a test for whether you have one, and SQL fixes, cheapest to strongest.
Haitao Wang|Oct 19, 2026