TVM Optimizations for edge computing

[amalik]

I am looking into DL optimizations for edge computing. Is there any work that looks into the TVM optimizations for edge computing?