Ok, try this. It's a bit long winded but;
1- hydrolyse the ester to and acid
2- derivatise the acid to and acyl chloride with thionyl chloride
3- attck the carbonyl with aziridine to form N-acetyl aziridine
4- react the alkyl lithium group you want for the other side of your ketone (in theory the grinard would work here too)
and voila! done. No tertiary alcohol in sight.
OR perhaps you could reduce your ester to and aldehyde and then just use the grinard. With no leaving group (H is a terrible leaving group) that should afford you a secondary alcohol that you could re oxidise to the desired ketone.
depends on your substrate