metadata
title: AIE WF2026 Qdrant Edge Hybrid Search
emoji: 🔎
colorFrom: indigo
colorTo: green
sdk: docker
app_port: 7860
pinned: false
license: apache-2.0
AI Engineer World's Fair 2026 — Qdrant Edge hybrid search
Search 560 WF2026 talks with a hybrid retriever running on embedded Qdrant Edge
(qdrant-edge-py, in-process) behind FastAPI — no cloud vector DB.
- Dense
bge-small-en-v1.5+ sparse BM25 (Qdrant/bm25, IDF) fused with RRF, all inside Qdrant Edge - Prebuilt index loaded from the GitHub release on boot; query embedded server-side
- Each result badged by retriever (Dense / BM25 / Both), ranked by RRF hybrid
Code: https://github.com/KShivendu/aie-wf2026-qdrant-edge-search