PomaiDB
pomagrenate/pomaidbThe predictable edge-native database for multimodal AI memory. Embedded, single-threaded, and offline-first with memory-mapped structures.
A comprehensive gallery of my systems-programming libraries, data-mining engines, local-first RAG components, and operational applications. All code is designed with a focus on predictability, low latency, and systems-level optimization.
System libraries, allocators, and specialized database engines focusing on resource predictability and memory bounds.
The predictable edge-native database for multimodal AI memory. Embedded, single-threaded, and offline-first with memory-mapped structures.
A hardware-aware DRAM bank partitioning memory allocator designed to protect edge systems from OOM failures and ensure strict performance isolation.
Cold storage and index archiving system optimized for flash wear-aware databases, serializing old vector embeddings dynamically.
Semantic cache layer optimized for LLM prompting to reduce redundant network and GPU cycles on repetitive tasks.
C++ pattern mining algorithms and high-throughput lexical scanners.
High-performance C++ Data Mining library implementing CHUO-Miner, MFHOI-Miner, HUPP-Miner, and oblivious shape-hiding tree-mining (VIFP).
A low-level C99 lexical analysis library featuring the Flat-Array Robin-Hood Offset Tokenizer (FARO-Tokenizer) for zero-dependency parsing.
Hybrid keyword and semantic query search engine using vector membranes and posting list structures.
Orchestration engines, context compressors, and autonomous agent loops.
Local-first Retrieval-Augmented Generation ecosystem coordinating C++ inference, embedded vectors, and Go autonomous cores.
The cognitive execution framework and local LLM agent execution loop for the Cheeserag autonomous cluster.
Utility aimed at optimizing token consumption for AI agents by identifying and compressing redundant context chunks.
AI-assisted semantic search engine combining hybrid sparse-dense representation and neural query expanders.
Operational prototypes and vertical solutions using embedded AI backends.
AI-assisted drug retrieval and patient management system showcasing description-based vector lookup.
Multi-workspace HR operations SaaS with embedded agent workflows for automated tasking, shifts, and wage computing.