As co-founder, Erik designed the semantic search engine, Earlybirdy. It delivers relevant results by parsing contextual data in real time. With its proprietary emotion algorithm and AI models, Earlybirdy predicts repeat purchase intent by revealing leading indicators of customer acquisition, retention, and churn—with 95% accuracy.
Earlybirdy was originally built on the Twitter API but can be used on any unstructured data.
Role
Co-founder, Product Designer
Dates
January, 2015 - June, 2015 • 6 months
Accomplishments
- Built a team of 6 developers specializing in A.I. and data science
- Seed-funded a startup from concept to MVP in 3 months
- Created a new approach to natural language processing (NLP)
- Designed an algorithm to find signal in noise in unstructured data
- Achieved a combined relevance and confidence of 95%
- Applied the algorithm to global public data at 1B posts per week
- Negotiated a 1-year global Twitter firehose partnership
- Built an automated net prompter score (NPS) for brands
- Inferred buying intent for user acquisition, retention, and churn
- Learned the risk of building a product on 3rd party data
