JJ657

Python, Qdrant, Milvus, PyTorch, FastEmbed, NumPy, Polars, DuckDB

Quantized vector indexing (Product Quantization), hybrid sparse-dense retrieval, and memory-mapped embedding stores.

Automated chunk boundary evaluators and lightweight local HNSW recall probes.

RAG retrieval latency audits, vector database benchmark discussions, and data preprocessing pipelines.