Content Catalogue Intelligence
Lifted content catalogue discoverability by 40% through programmatic metadata enrichment for 22M+ content assets.
The Challenge
JioTV+'s vast content catalogue of 22M+ assets suffered from inconsistent and sparse metadata. Poor entity recognition meant users couldn't discover content through search or recommendations, reducing the effective catalogue size significantly.
The Solution
- Designed a programmatic metadata enrichment pipeline processing 22M+ content assets
- Optimized entity recognition across the entire catalogue for better search and recommendation signals
- Built automated pipelines that continuously enrich new content as it's added to the catalogue
- Improved the quality signals feeding into search ranking and recommendation algorithms
Impact
40%
Discoverability Lift
22M+
Content Assets
Key Decisions
- Programmatic over manual: Automated enrichment scales with the catalogue while manual tagging creates permanent operational cost.
- Search-first optimization: Prioritized metadata fields that directly impact search retrieval and recommendation quality over general completeness.
Tech Stack
Python
SQL
NLP
Entity Recognition
REST APIs