There is no point in having a machine controller without having machines for it to control.
About this creation
During one of many periods in which mechanical combatants were unpopular in Imperial military circles, an arms company designed a prototype series of advanced, automated machineries of destruction. Of these, one was chosen by Imperial scientists for study and tests. That was the PAM-4 assault robot. The type was self aware, and fully capable of commanding itself. However, intelligent machines are not wise things to have, and the Empire retained only one, with its mechanical thought processor disabled, only able to be brought back to life by a command code known only to the Emperor and certain trusted members of his staff.
Several years later, a small rebellion occured. Imperial Guardsmen refused to fire apon the weaponless crowd, and retreated. After their court-martial and execution, it was decided that if a combat robot was a part of the squad in question, the refusal of the troops to fire would have been overridden by the cold and uncaring mind of the machine warrior and the fiasco would not have occured.
The PAM-4 was revived and its brain removed, to be replaced by a less advanced proccessor only capable of following orders. This was an even worse fiasco, as the prototypes were unable to differentiate between Imperial infantry and the forces of the vile foe. Yet worse, the inflexible mind was incapable of adjusting for changes in the environment, attempting to climb up stairs that it had ordered the small machine-minds of its missiles to destroy only seconds before.
That sort of thing was fine for turrets, which are immobile, but for a machine intended to murder officers and commanders as well as foot troops, it was not a good thing.
The problem was easily fixed, however, by the inclusion of new programming and over-ride circuits keyed to the new C-995 command unit. The automated commander was intelligent enough to think as well as an average human, and it was focused on one thing only: to destroy and wage tactical warfare apon the designated enemy. The sharing of minds allowed the new T-402 to perform as well as if it was intelligent iteself, and there was no chance of rebellion; the commanders were set with remote destruct charges in case of revolt.
Its vision was a standard Zenarms monocular optical sensor, capable of seeing as well at night and in fog as in the light of day. It was fast, agile, tough and well armed, and it was able to communicate with its commander at immense distances.
It's left arm was taken up entirely by a ZERA-229 60mm missile pod; twin tubes, long range, hybrid flechette/explosive fragmentation/armour piercing ammunition.
The right was a 27mm VERA light autocannon, firing explosive fragmentation AP rounds and ejecting to the top, with a firing rate of 450 rounds per minute. This was mounted at the top of the arm, and the bottom section was taken by a FUR-3 2cm airburst anti-personnell grenade launcher with a practical rate of fire of 300 rpm.
It could take to the skies.
Not well, mind you, as it could only fire its boosters at a 45 degree angle without a suitable item to rest on prior to launch, but it could use its claws for purposes other than digging and looking good during flight.
Should I mention that there was a big stink from human rights groups about that? Nah, why bother.
Of course, it wasn't very popular when overflying infantry formations, and it was difficult to manually control during flight, but it was a useful ability, especially so since Zenexabas commanders usually left small forces of machine warriors behind during withdrawals to act as automated guerillas until they could return. Sometimes they never came back, and the machines stayed behind, and sometimes the Empire simply forgot about them and they stayed to kill Imperial citizens for centuries to come.