My bad, was improper use of signatures, should've used np.ndarray instead of np.array
Leaving here in case anyone searches
We are testing flyte as a potential solution to replace an inhouse developed solution.
Our data is often multi dimensional (3+ dimensions) and large (each dimension can be pretty large). Our code represents the data with ndarrays.
How do you handle encoding and decoding of ndarrays to do so efficiently?