Credential-building open issues here — a merged pull request is verifiable proof of work for your résumé.
Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2
Real code, on a real project, reviewed by a real maintainer — a merged PR here is credential-building proof of work you can point to.
Computed from your public GitHub only — nothing is stored.
Connect GitHub to check →openinfer-project/openinfer · issue #780 · opened Jul 2026 · no PRs referenced when indexed
Under the PerToken batch-invariance policy, each bucket-N decode CUDA graph bakes ~36 layers × 6 projections × N cublasGemmEx nodes, because the PerToken GEMM…
Click Claim →on any issue above to open its claim page, where you'll get a terminalhire claim <token> command and a one-tap deep link into your terminal.
terminalhire contribute