B2B software company · 2025
LLM-Powered Semantic Search
Semantic search over a large content catalog using embeddings, improving relevance and discovery over keyword search.
- Client
- B2B software company
- Role
- Project Manager
- Timeline
- Jun 2025 – Oct 2025
- Domain
- AI / Search
- Approach
- Agile
- Team size
- 5–9
- Budget
- Up to $250K
Objective
Deliver semantic search over a large content catalog, improving relevance and discovery beyond keyword matching.
What I did
- ▸ Defined search relevance goals and evaluation with stakeholders
- ▸ Coordinated embedding pipeline and index design with engineering
- ▸ Planned rollout and A/B evaluation
- ▸ Tracked relevance metrics and iterated
- ▸ Coordinated UI integration of search results
Key deliverables
- ▸ Embedding & indexing pipeline
- ▸ Semantic search API
- ▸ Relevance evaluation
- ▸ Search UI integration
- ▸ Production rollout
Outcome
Improved search relevance and content discovery compared to the previous keyword-based search.
Tools & tech
- Embeddings
- Vector DB
- Search APIs
- Python

