Kronos
The metadata engine behind content discovery on Titan OS. Kronos collects content catalogs from streaming partners, cleans them up and enriches them. Search, recommendations and the home screen all work from the same data.
- Role
- Product Manager
- Partner Catalogs
- 25+
- Titles Processed
- 1M+
- Markets Served
- 40+
A TV home screen is only as good as the data behind it. If a title shows up twice, has no artwork or sits in the wrong genre, viewers notice right away. At Titan OS, Kronos prevents that. It takes in catalogs from global and regional streaming partners, each with its own format and level of quality, and turns them into one consistent set of records.
As Product Manager for Metadata, I own the product strategy and roadmap for Kronos. That includes how partner catalogs are ingested, how we match the same title across providers, where AI fills the gaps, and how the data reaches search, recommendations and the content rows on the home screen.
Features in Detail
- Catalog ingestion: Takes in metadata feeds from global streaming services and local partners and maps them to one shared data model.
- Entity resolution: Recognizes when different providers deliver the same movie, series or episode and links them, so each title shows up only once.
- AI enrichment: Uses machine learning to add missing tags, fix genre and taxonomy assignments and make records more complete.
- Search and recommendations: Delivers structured data fast enough for search and the recommendation engine to use directly.
- Content rows: Drives the curated, regional and personalized rows on the home screen.
- Quality checks: Validation rules and monitoring catch broken or incomplete data before it reaches viewers TVs.
My Role
- Strategy: I set the long-term vision for Kronos, define the metadata requirements and own the roadmap across teams.
- Discovery and delivery: I run discovery with internal platform teams and content partners, lead sprint execution and bring AI initiatives into the product.
What’s Next
- Neural networks for matching: Using neural networks to normalize partner data and match titles across providers more accurately.
- Better personalization: Using regional viewing trends and user information to make discovery rows more relevant in real time.
- Central content hub: One place for all metadata across the company, so search, ads and the editorial team all work from the same data.
In-house metadata platform, built for Titan OS.