Three principles for building a vector platform at Thumbtack

Summary

Thumbtack describes a vector platform that lets ML engineers stand up production vector search without separately negotiating database access, building ETL, or writing a query service. The design uses an existing Aurora Postgres cluster with pgvector, while Airflow jobs generate embeddings into BigQuery and managed reverse-ETL syncs them into vector tables; a query API validates inputs and returns ranked results. Three principles guide the layers: reuse trusted infrastructure, treat embeddings as ordinary data flowing through established pipelines, and lower adoption costs for subsequent teams. Production examples include semantic customer-support escalation using cosine similarity and an internal knowledge assistant combining vector retrieval, often with keyword search, before a language model responds; planned directions include self-serve provisioning, vector-backed feature retrieval, and framework-neutral agent memory.

Takeaways

1. 1

   pgvector lets the platform use existing Postgres operational practices, including schema migrations, access roles, backups, replicas, monitoring, and query investigation.

2. 2

   Embedding refreshes use Airflow, BigQuery, and managed reverse-ETL, inheriting established scheduling, retry, schema-mapping, alerting, and observability patterns.

3. 3

   Current uses include semantic support escalation and knowledge retrieval; planned uses include vector-backed feature retrieval and framework-neutral memory for agentic workflows.