[Python] Remote execution speed up doesn't support custom operations

Required prerequisites

• Consult the security policy. If reporting a security vulnerability, do not report the bug using this form. Use the process described in the policy to report the issue.
• Make sure you've read the documentation. Your issue may be addressed there.
• Search the issue tracker to verify that this hasn't already been reported. +1 or comment there if it has.
• If possible, make a PR with a failing test to give us a starting point to work on!

Describe the bug

E       RuntimeError: Failed to launch kernel. Error: Failed to execute the kernel on the remote server: "Failed to process incoming request"
E       Error message: "test_remote_code_exec.py:364: error: unhandled function call - custom_h, known kernels are dict_keys([])\n\t (offending source -> custom_h(qubits[0]))"

Steps to reproduce the bug

def test_arbitrary_unitary_synthesis():
    cudaq.register_operation("custom_h",
                             1. / np.sqrt(2.) * np.array([1, 1, 1, -1]))
    cudaq.register_operation("custom_x", np.array([0, 1, 1, 0]))

    @cudaq.kernel
    def bell(angles: list[float]):
        qubits = cudaq.qvector(2)
        custom_h(qubits[0])
        custom_x.ctrl(qubits[0], qubits[1])
        ry(angles[0], qubits[1])

    hamiltonian = 5.907 - 2.1433 * spin.x(0) * spin.x(1) - 2.1433 * spin.y(
        0) * spin.y(1) + .21829 * spin.z(0) - 6.125 * spin.z(1)

    optimizer = cudaq.optimizers.Adam()
    energy, parameter = cudaq.vqe(kernel=bell,
                                  spin_operator=hamiltonian,
                                  optimizer=optimizer,
                                  parameter_count=1)
    print(f"\nminimized <H> = {round(energy,16)}")
    print(f"optimal theta = {round(parameter[0],16)}")

Expected behavior

Application program runs successfully.

Is this a regression? If it is, put the last known working version (or commit) here.

Not a regression

Environment

• CUDA Quantum version: main (or PR#1794 if not yet merged)
• Python version: any
• C++ compiler: N/A
• Operating system:

Suggestions

No response

Start with the failing reproduction in test_remote_code_exec.py at line 364, especially test_arbitrary_unitary_synthesis and its cudaq.register_operation calls. Run the remote execution case and trace how custom_h and custom_x are handled by the remote server. Done means the application runs successfully with custom operations, without the reported kernel-launch error.