Engineering projects

kaggle-vllm

kaggle-vllm is the primary project documented here. It is a compatibility and runtime-delivery toolkit for running upstream vLLM in the explicitly validated Kaggle dual-NVIDIA-T4 environment. Its public repository separates runtime provenance, compatibility checks, experiment protocols, and evidence artifacts from narrative claims.

• Source repository
• Documentation
• Research evidence
• Release history

The project wraps and validates upstream vLLM; it is not a fork or a replacement inference engine.

Publication queue

The following repositories are reserved for future project notes after their authorship, current status, and evidence have been reviewed. Listing them here does not assert original authorship or active maintenance.

• llamatelemetry
• CommGuard
• selected Edge-Computing-LLM work
• Ubuntu-Cuda-Llama.cpp-Executable

Descriptions will be added only with repository-specific evidence.