Articles

Building a Semantic Search Engine for arXiv with Fine-Tuned Embeddings

retrieval | indexing | embedding | research
Search over scientific papers gets difficult once the query stops looking like a title fragment and starts looking like an intent: a method, a problem, or...

AgenticKernel: 43× C++ Kernel Speedup with LLMs

llm | performance optimization | C++ | agentic workflows
AutoResearch showed that a simple agentic feedback loop can iteratively improve PyTorch models: propose a change, evaluate the result, and repeat. Compute...

Speeding Up Binary Feature Matching with a Cache-Friendly Hash Join

Computer Vision | Search | Indexing | Performance Optimization | C++
Feature matching is a core operation in the hot loop of visual odometry, SLAM, and structure-from-motion. Classical pipelines treat it as a nearest-neighbor...