Articles
Engineering write-ups on durable execution, reliability, agentic systems, security, and infrastructure, from the team building Vectorbea.
9 articles
Lessons from Building Vectorbea v1
What we'd keep and what we'd change across UI, backend, security, observability, and positioning, after shipping the first version of Vectorbea's durable workflow engine.
Self-Correction Loops for Failed Workflows: Blind Retry Isn't Intelligence
The difference between retrying a failed step and helping a workflow understand why it failed, error classification, bounded self-correction, and where we draw the line and call a human.
Cost Budgets and Rate Limits for Agentic Workflows
How we estimate token costs before and during a run, enforce per-run and per-workspace budgets, apply rate limits, and build kill switches that actually stop a runaway workflow.
Worker Scaling with Redis Streams: Consumer Groups, PEL, and When to Reach for Kafka
How Vectorbea's worker fleet pulls work from Redis Streams, consumer groups, the pending entries list, retry and DLQ handling, and the honest answer to 'why not Kafka?'
BYOK Architecture for an AI SaaS: Benefits, Risks, and Boundaries
Why we let customers bring their own LLM provider keys, what it costs them and us, and the security boundaries we think any BYOK system needs, without the implementation specifics.
Human Approval Gates in Agentic Systems
Modeling 'wait for a human' as a first-class workflow step, waiting states, timeouts, escalation, and why the audit trail has to be airtight.
Retries, Resume, and Idempotency: The Unglamorous Core of Reliability
Step-level retries sound simple until you ask what happens when the step already had a side effect. Notes on idempotency keys, resumption, and duplicate tool calls in Vectorbea.
Designing Event History as a Primitive for AI Workflows
How we modeled the append-only event log that backs every Vectorbea run, and why we treat it as the source of truth rather than an audit trail bolted on afterward.
Why Long-Running AI Workflows Need Durable Execution
Async jobs and retry decorators get you most of the way to a working agent, and then they don't. Here's why we built Vectorbea around durable execution from day one.