A high-performance inference system for large language models, designed for production environments.
-
Updated
Dec 19, 2025 - C++
A high-performance inference system for large language models, designed for production environments.
Draft-Target Disaggregation LLM Serving System via Parallel Speculative Decoding.
An out-of-order execution CPU simulator for CS2410 Computer Architecture course final project at the University of Pittsburgh.
An adaptive, speculative request-hedging library that automatically cuts p95 tail latency
The USOURCE Framework: Natural Language → Universal Source Code. One translation problem. One hypothesis. Testable predictions or it's just a story.
Piglog partitions and reassembles partitions from Detlog
Speculative branching cache engine with optimistic versioned commits. Inspired by MVCC, OCC and local-first architectures. Includes design doc, testing strategy, benchmarks and HTTP API.
Bio World Wide Web – Generative Climate Organism. "A living web organism that metabolizes Indonesian climate data into fungal-neural-plant tissue, continuously growing, decaying, and regenerating without endpoint."
Source code for a speculative prototype object that explores anti-interface technology.
Hello! I am silviciouƨ, a fashion futurist specialised in digital fashion. Come in!
Fictional rail transit draft for the Yanji metro area—metro, intercity & HSR maps in a speculative Northeast Asia setting.
Otis MFAGD_Seminar Studio III_Spring 2019
🧠 Generate wikis from code repositories with Pearl, a RAG system in Elixir and Phoenix, allowing natural language querying of any codebase.
To associate your repository with the speculative topic, visit your repo's landing page and select "manage topics."