Search a Name Across University Databases

Problem Design an approach to search for a given student's name across the records of every university in India, and explain how you would optimise it for speed at that scale.

Input / Output

• Input: a query name (possibly partial or misspelled), against a corpus of student records spread across many universities.
• Output: all records matching that name, ideally ranked by match quality.

Constraints

• The corpus is very large — tens of millions of records across thousands of institutions — so it will not sit comfortably on one machine.
• The workload is read-heavy with repeated queries; writes (new enrolments) are rare and arrive in batches.
• Names are messy: case differences, extra whitespace, initials, transliteration variants and outright typos are all common.
• A full scan per query is unacceptable at this scale.

Example

• The query "ravi kumar" should match records stored as "Ravi Kumar", "RAVI KUMAR" and "Ravi Kumar", and ideally "Ravi Kumaar" under fuzzy matching, regardless of which university holds each record.