numpy built with locally built blis does not use multithreading

I'm looking for help with an issue I'm having building Numpy against locally built blis for zen3.
I've configured blis to enable threading using openmp. (it is installed and working on my machine, validates using a cpp code).

After building blis and numpy against it, it seems like numpy is only using single thread.
I'm not able to find out the reason.

Here is the showconfig output:

configuration family:       zen3
sub-configurations:         zen3
requisite kernels sets:     zen3 zen2 zen haswell
kernel-to-config map:       haswell:zen3 zen:zen3 zen2:zen3 zen3:zen3
-------------------------
BLIS version string:        0.9.0-118
.so major version:          4
.so minor.build vers:       0.0
install libdir:             /home/me/blis/lib
install includedir:         /home/me/blis/include
install sharedir:           /home/me/blis/share
debugging status:           off
enable AddressSanitizer?    no
enabled threading model(s): openmp single
enable BLAS API?            yes
enable CBLAS API?           yes
build static library?       yes
build shared library?       yes
ARG_MAX hack enabled?       no

I have set OMP_NUM_THREADS=64 (using 64 threads cpu) and export BLIS_THREAD_IMPL=openmp just in case.

Using pip's version of numpy, multhreading works (but it's configured for OpenBLAS). numpy's config shows it is configured to use blis:

    libraries = ['blis', 'blis']
    library_dirs = ['/home/me/blis/lib']
    define_macros = [('HAVE_CBLAS', None)]
    include_dirs = ['/home/me/blis/include/blis']
    language = c
    runtime_library_dirs = ['/home/or/blis/lib']
blas_opt_info:
    libraries = ['blis', 'blis']
    library_dirs = ['/home/me/blis/lib']
    define_macros = [('HAVE_CBLAS', None)]
    include_dirs = ['/home/me/blis/include/blis']
    language = c
    runtime_library_dirs = ['/home/me/blis/lib']