QDOX RESEARCH
From a question
to evidence.
Edge computing, efficient inference and physical robotics. We share the work at its actual stage, with enough context to understand what the evidence does—and does not—show.
TECHNICAL REPORTS
Research papers
Technical notes and reviews built from public sources, with numbered equations, diagrams and references. Each paper states whether it examines recorded evidence or proposes a method for future testing.
Efficient Language-Model Inference at the Edge
Public research on memory, heterogeneous scheduling and speculative decoding, with equations and a proposed evaluation method.
Humanoid Model and Balance Validation
Mass consistency, reduced balance equations and a proposed path from analytical checks to physical evidence.
Contact Verification for Legged Robots
Contact models, numerical diagnostics and a proposed protocol for evaluating simulation evidence.
Public Evidence for Edge Compute on NVIDIA Jetson Thor
What public compiler probes and device records establish—and what still requires runtime experiments.
Computing foundations
The Thor note examines public compiler probes. The inference review explains memory, scheduling and latency from published literature. Fast-LLM is our open-source C++17 inference runtime, released under MIT and still in development.
Explore Thor on GitHub ↗Explore Fast-LLM on GitHub ↗
Physical foundations
The quadruped is in design. The humanoid is in the pipeline. These public-literature reviews explain contact, balance and validation methods; they report no QDOX robot experiments.