04
Search and ranking for startup–corporate matching
GlassDollar · Data Engineer · 4 months
The challenge
GlassDollar connects large corporates with the right startups. Across hundreds of thousands of companies, search results are only valuable when the data is reliable and the most relevant matches surface first.
The approach
- 01
Ranking algorithms and data pipelines that improve relevance and discoverability across the platform.
- 02
Scalable data collection with scrapers (Firecrawl, Exa, Crawl4AI) and enrichment APIs (Clearbit, LinkedIn, Crunchbase).
- 03
Data quality through deduplication, benchmarking and usage analytics (PostHog).
- 04
Full-stack delivery with TypeScript/Node.js for search services and APIs, and Python for embeddings, clustering and evaluation.
The outcome
Production-ready search features with improved relevance across 240k+ partnerships and 2k+ corporates.
My role
Data engineer working across the stack, in close collaboration with product and engineering teams.