[GSoC 2025] Beam ML Integration Tracking: Vector DB, Feature Store, and Embedding Generator

[mohamedawnallah]

Upstream

🔗 Apache JIRA: GSOC-279 🔗 GSoC 2025 - Project Page

Mentor

Mentor: Danny McCormick GitHub Handle: @damccorm

Legend

• ⬜: Not started yet
• ⚪: Feedback — not yet addressed
• 🏗️: Created
• ⏳: In review
• ✅: Merged

Milestone 0 – Community Engagement

Status	Period	Activity	Role	Artifact
🏗️	Community Bonding	Share GSoC 2025 selection and project goals with Beam developer community	Author	Proposal Mail
⏳	Community Bonding	Enrich data with CloudSQL in the Python SDK	Author	PR #34398

Milestone 1 — Design Doc for Milvus Vector DB Enrichment Handler Integration

Status	Period	Activity	Role	Artifact
🏗️	May 26 - May 30	Write design document for Milvus enrichment handler	Author	Proposal
🏗️	May 26 - May 30	Share design doc on Beam dev mailing list for feedback	Communicator	Mail

Milestone 2 – Implementation of Milvus Enrichment Handler

Status	Period	Activity	Role	Artifact
⏳	May 31 - Jun 13	Implement MilvusEnrichmentHandler in ml/rag/enrichment	Author	PR #35216
⏳	May 31 - Jun 13	Unit test MilvusEnrichmentHandler	Author	PR #35216
⏳	May 31 - Jun 13	Integrate test MilvusEnrichmentHandler	Author	PR #35216
⏳	May 31 - Jun 13	Update enrichment handlers documentation	Author	PR #35216
⏳	May 31 - Jun 13	Add milvus_enrichment_transform.ipynb notebook example	Author	PR #35216

Milestone 3 – Design Doc for Milvus Sink I/O

Milestone 4 – Implementation of Milvus Sink I/O

Milestone 5 – Design Doc for OpenAI Embedding Generator

Milestone 6 – Implementation for OpenAI Embedding Generator

Milestone 7 – Design Doc for Feast Feature Store Enrichment Handler

Milestone 8 – Implementation of Feast Feature Store Enrichment Handler

Milestone 9 – Design Doc for Feast Feature Store Sink I/O

Milestone 10 – Implementation of Feast Feature Store Sink I/O

Nice to Have