Mini-project Retail, E-commerce & Consumer Goods Case Studio

Spelling and sound-alike fixer for product search

Inspired by publicly reported work at Flipkart

Correct misspelt and sound-alike product searches typed by Indian shoppers, so they still find the product.

The brief

Flipkart's tech blog has written about search that copes with Indian spellings and phonetics, where shoppers type words as they sound. Build a query corrector: generate realistic misspellings of product queries, then correct them with edit distance, phonetic matching and a small model, and measure how many corrected queries find the right products. Good looks like: a test set of real-looking errors, accuracy for each method, and a demo search box.

Read the case first

What Flipkart has said in public about this kind of work, on its own pages:

Make it your own

Add a language: Telugu, Hindi or Tamil words typed in English letters for groceries, clothes or medicines. Collect 300 real queries from friends and family (no personal details) as your test set.

AITeky has no link with Flipkart, which has not reviewed or endorsed this brief. The brief only draws on what the company has published, uses open data rather than its data, and is a learning exercise: don't present your project as Flipkart's work or system.

What you will build

Collect product names and queries from open catalogue data
Generate realistic misspellings and sound-alike variants
Correct with edit distance and phonetic keys
Train a small character-level model
Compare methods on the test set
Build a search box demo

Free, open data to build it with

It takes a minute with your email. AiTeky Pass (₹999/month or ₹9,999/year) is free for your first 30 days.