Blog

AI agents, RAG, and integrations — notes from production.

What to Put in a Product Embedding: Composing the Text You Index

Product embedding text should be a clean composed field - title, key attributes, category, short description - not raw HTML or boilerplate.

The Async Re-Embed Pipeline: Keeping a Vector Index Fresh Without Blocking Writes

An async re-embedding pipeline: a write trigger, a filter for semantic-field changes, and a queue worker - so catalog updates never block on an embedding call.

Tuning Top-N: How Many Candidates Should Go Into the Re-Ranker?

Re-ranker top-N tuning: too small and the re-ranker can only reorder a weak set; too large and you pay latency and cost on noise. Tune against recall at N.

Choosing an Embedding Model for Product Search: Small, Large, or Multilingual

An embedding model for product search: small is the default for an English catalog at scale, large adds a few points at ~5x cost, multilingual for non-English.

Handling Angry Customers: When an AI Support Agent Should Not Try

With frustrated customers, an AI support agent should acknowledge once and escalate fast - trying to de-escalate reads as being fobbed off and delays the fix.

Cutting Over From a SaaS Support Bot to a Custom Agent Without a Gap

To migrate a support bot to a custom agent, run both in parallel, cut over one ticket category at a time, and keep a routing flag as the rollback.

Computer Vision for Fresh and Perishable Sections: Where It Gets Hard

Computer vision for fresh produce retail works at bay level, not item level - no consistent packaging, variable appearance, fuzzy stockouts, and fast turnover.

Detecting Misplaced Products and Planogram Deviation, Not Just Gaps

Planogram compliance computer vision flags misplaced products and section-level deviation - it needs a machine-readable planogram and reliable product ID.

Sizing and Staffing the Human Review Queue for Ambiguous Detections

The computer vision human review queue takes the ambiguous confidence band - size it by detections times band width, and keep each review to seconds.

MQTT Quality of Service: Getting a Detection Event Out of a Flaky Store

MQTT QoS for retail edge: use QoS 1 (at-least-once) plus downstream dedupe for stockout alerts - QoS 0 can drop them, QoS 2 is overkill for idempotent events.