The fraction of women at the higher levels in Australian STEMM remains at 20-30%, despite having achieved parity at level A (junior postdoc) for nearly two decades.  I present an evidence-based Monte Carlo model that can reproduce the current gender disparity at all levels in Australian science, starting with the 2002 data.  I show that a small 2-3% implicit bias in recruitment and/or promotion can explain the current gender gap.  To provide institutions with guidance in closing the gap, I run my models forward in time, while including various initiatives, such as 50:50 hiring, a goal of 50:50 departures, and equal retirements.  I show that the current gender gap for levels B-E in Australian STEMM in academia can be closed within 12-15 years with the implementation of reasonable and fair initiatives.  These results have important implications for future workforce planning at institutions worldwide.