Evaluate GPU kernel tasks for technical accuracy, realism, solvability, reproducibility, and robust testing criteria.
Rigorously test and troubleshoot complex GPU programming scenarios to identify memory allocation bugs, execution bottlenecks, and parallel computing logic errors.
Review CUDA, Triton, and other GPU kernel implementations and provide clear, actionable technical feedback.
Our partner is a company specializing in AI training and evaluation, seeking experienced GPU kernel specialists to audit AI training tasks. The project is globally distributed and offers fully remote freelance work.
Build reference architectures, benchmarks, and documentation that engineers trust.
Own the developer community, answering hard questions and setting the tone.
Write production-grade code for integrations and tooling that lower the barrier for new users.
Andromeda Cluster gives early-stage startups access to scaled AI infrastructure once reserved for hyperscalers. We are a unicorn with a small senior team, building the liquidity layer for global AI compute.
Own the cost and performance of the inference stack, improving throughput and latency without compromising reliability.
Optimize through KV-cache management, continuous batching, speculative decoding, and quantization.
Work within serving engines like vLLM, SGLang, and TensorRT-LLM, profiling performance down to kernel level.
Adaption builds efficient AI that evolves in real-time, making intelligence flexible, personalized, and accessible to everyone. They focus on talent density, bringing together driven individuals to push the boundaries of continual adaptation.