It was actually built by the Ioffe Physico-Technical Institute in St. Petersburg.
I know these guys. They are serious, and know their stuff well.
However, their stuff is MWIR photodiodes and LEDs - an IR band somewhat above thermal IR.
I got one of their LEDs and a couple of photodiodes of 3.4um wavelength, for a camera project.
Here is an english datasheet -
http://www.ioffeled.com/Specifications/LED34.pdf (3.4um, theirs was said to be 2.4um, but i can't find such one).
The interesting thing about that LED is that it runs at very low forward voltage - under 0.2V, and less that 0.1V at the temperature specified.
Also, at 135*C the device itself will be emitting some radiation in it's own wavelength by blackbody heat.
So, it can run on raw output of a termocouple, or there can be some peltier-style effects of the heat gradient being re-absorbed in the LED material itself.
Would be nice to read the full MIT paper.
EDIT: Noticed the link to the paper.
They did have the problem i mentioned:
Since the photodetector remained unheated (~ 25*C), even without bias the LED’s 135*C active region emits approximately 40 nW of blackbody radiation in this wavelength range; the lock-in technique was necessary to separate the 69 pW of electrically driven optical power from the portion of that blackbody background which was incident on the detector surface.