Edge Recommender

“More like this” recommendations running on store or app hardware.

Industry	Level	Difficulty	Recommended mode
Retail & e-commerce	5 of 5	Frontier	sq8

Coming soon

A complete, runnable Edge Recommender reference implementation is in progress. The architecture and core code below already work with SrvDB 0.3.

The problem

• Recommendation services are costly.
• Cold-start items get no exposure.
• Latency matters at the point of sale.

How SrvDB powers Edge Recommender

1. Item embeddings from content (text, image) solve cold start.
2. In-process search keeps latency low.
3. Compressed storage with sq8.
4. Rust or Python back-ends.

What changes

Today	With Edge Recommender on SrvDB
Heavy recommendation infra	A library call
Cold-start invisibility	Content-based similarity
Network latency	In-process

How it works

1. Embed items.
2. Store with sq8.
3. Recommend by item vector.
4. A/B test.
5. Refresh.

SrvDB features it uses

• SQ8 compression (sq8): stores each value in one byte, 4× smaller than full precision.
• search_batch: runs many queries at once, in parallel inside Rust.
• Rust API: embed SrvDB directly in Rust services and devices.
• Crash-safe persist(): after it returns, data survives a crash or power loss and reopens exactly as committed.

Use cases

• More like this
• Point-of-sale suggestions
• Email recommendations
• Cold-start exposure
• Trend detection
• Bundles

Core code

embed() is your embedding model (for example sentence-transformers, or a function that turns your records into feature vectors).

import json
import srvdb

db = srvdb.Database("./edge-recommender-db", dimension=256, mode="sq8")

def ingest(items):
    """items: list of (id, text, metadata dict)"""
    db.upsert(
        ids=[i for i, _, _ in items],
        embeddings=[embed(t) for _, t, _ in items],
        metadatas=[json.dumps(m) for _, _, m in items],
    )
    db.persist()                      # crash-safe checkpoint

def more_like_this(item_description, k=5):
    hits = db.search(embed(item_description), k)
    return [(json.loads(db.get(i)), round(score, 3)) for i, score in hits]

Limits and honest notes

• Benchmark latency on your hardware.

Build roadmap

1. Prototype: load a sample of real data with flat mode and check that searches return what an expert expects.
2. Measure: if you move to an approximate mode, compare its results with flat on the same queries (recall).
3. Pipeline: add continuous upsert and decide when to call persist().
4. Harden: test restarts and crashes, add logging, review every output path with domain experts.
5. Deploy: ship the database directory with the application or service.

All Retail & e-commerce projects

Level	Project	What it does
1 · Beginner	Semantic Product Search	Shoppers describe what they want and find it.
2 · Intermediate	Shopping Assistant	Answers product questions from your catalogue and reviews.
3 · Advanced	Listing Guardian	Detects duplicate and counterfeit-looking listings.
4 · Expert	Multimodal Kiosk	In-store kiosk that searches by text or photo.
5 · Frontier	Edge Recommender (this page)	“More like this” recommendations running on store or app hardware.

See also: Use cases · Live examples (coming soon)