Show notes
How to reduce AI inference costs. This has become just as pressing a challenge for ML engineers today as model performance itself. A compiler is originally a tool that converts a program into code that machines can execute at maximum speed. However, traditional compilers operate on the premise that once optimization is done, it's finished. AI inference workloads don't work that way. Model versions change, batch sizes fluctuate, and usage patterns shift constantly. Statically optimized kernels quickly become suboptimal. What Together AI's TKCC, Together Compiler, is trying to solve is…
Website: tech.airadio.news

