Update mccall_model.md: Convert from Numba to JAX implementation

Converted the McCall job search model lecture from using Numba to JAX for better performance and modern functional programming approach.

Key changes:
- Replaced Numba's @jit and @jitclass with JAX's @jax.jit
- Converted NumPy arrays to JAX arrays (jnp)
- Used NamedTuple instead of jitclass for model classes
- Implemented JAX-style while_loop for iterations
- Added vmap for efficient vectorized computations
- Updated all code examples and exercises to use JAX

The conversion maintains all functionality while providing improved performance and compatibility with modern Python scientific computing stack.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <[email protected]>

Author: jstac