Clearly is isn't as you never put the input signal across the two inputs (pins 2 & 3) with an opamp (unless its a transconductance amplifier and you have a current source).
What you've done cannot work as the two inputs are a virtually short (this means the feedback network works to keep the two inputs at the exact same voltage). Given the thermocouple is itself an extremely low-impedance source this means the op-amp saturates as it fails to achieve negative-feedback lock.
Your input voltage goes between GND and the non-inverting input (pin 3) and the feedback network is as you describe on the inverting input (pin 2)
You realize the LM423 has an input offset error voltage of 7000uV - this will dwarf the thermocouple output.
Also thermocouple circuits need cold-junction compensation since the voltage depends on the temperatures at both ends of the circuit where dissimilar metals meet.