Jeff Hajewski
Jeff Hajewski is an individual software developer working on AI infrastructure and systems software. His principal active project, LatticeDB, is an embedded, single-file property-graph database that combines relationship traversal, vector similarity search, BM25 full-text search, durable streams, and ACID transactions in one local engine and query layer.
At a Glance
- AI and machine-learning application developers
- Developers building RAG systems and AI agents
- Teams creating knowledge graphs and local knowledge tools
- Developers needing embedded, single-machine graph/vector/text retrieval
- +1 more
AI Tools by Jeff Hajewski
(1)
LatticeDB
Embedded Graph Vector Search DB
Discussions
No discussions yet
Be the first to start a discussion about Jeff Hajewski
Latest News
Products & Services
MIT-licensed embedded single-file property-graph database for local applications, with graph traversal, HNSW vector search, BM25 full-text search, durable streams/changefeeds, and ACID transactions.
Python package exposing the database, Cypher queries, graph transactions, vector search, BM25 and fuzzy full-text search, embeddings, property indexes, durable streams, and changefeeds. PyPI shows version 0.14.0.
Node.js/TypeScript bindings with bundled type declarations and APIs for graph operations, vector and full-text search, transactions, embeddings, property indexes, streams, and changefeeds. npm shows version 0.14.0.
Go bindings using cgo, with an installed pkg-config workflow and an in-repository development workflow; the repository includes a graph/vector/text retrieval example.
Market Position
LatticeDB positions itself as the SQLite-like embedded option for connected data: unlike server-oriented Neo4j, Weaviate, Qdrant, and Pinecone, it is a zero-configuration single-file local engine; unlike SQLite, it natively combines graph traversal, HNSW vector search, and BM25 full-text search in one query layer. Its own comparisons also discuss Kùzu/LadybugDB, FAISS, pgvector, and Chroma, while acknowledging that alternatives may be stronger for server-scale deployment or specialized workloads.
Leadership
Founders
Jeff Hajewski
Creator and maintainer of LatticeDB and related projects. He earned a PhD in Computer Science from the University of Iowa in 2020 under Suely Oliveira; his doctoral research focused on efficient machine learning across many computers and evolutionary neural-network design. After the PhD he worked at Salesforce, Noom, and SAP, building distributed machine-learning systems, a machine-learning platform, and generative-AI systems respectively; his personal site also identifies prior work at Google and Citrix.
Executive Team
Jeff Hajewski
Creator and maintainer; CX AI Engineer at SAP
University of Iowa PhD in Computer Science (2020); previous work at Salesforce and Noom on distributed machine learning and ML platforms, with Google and Citrix also listed in his professional history.
Founding Story
LatticeDB began as Jeff Hajewski's open-source effort to provide a local, zero-configuration alternative for connected data: a single portable file and one query layer for graph relationships, semantic/vector retrieval, and text search. The repository's initial commit and MIT license copyright are dated December 2025.
Target Markets
- AI and machine-learning application developers
- Developers building RAG systems and AI agents
- Teams creating knowledge graphs and local knowledge tools
- Developers needing embedded, single-machine graph/vector/text retrieval
- Python, TypeScript/Node.js, Go, C, and CLI users
- Graph RAG and retrieval systems
- AI-agent persistent memory
- Knowledge graphs and linked notes/documents
- Local knowledge tools and single-machine applications
- Local development and prototyping as a lightweight alternative to server databases
- Relationship-heavy workloads requiring graph, semantic, and textual retrieval together