The paper presents the assembling flux of thermocouple-instrumented nuclear fuel element for research reactor, from the point of view of the welding / brazing engineer, considering nuclear quality and safety requirements: fuel element structural reliability (no radioactive leaks through joints) and temperature signal reliability (thermocouple sheath integrity), this signal being an essential parameter for reactor normal operation and emergency shut-down. The paper is a real case study for an experimental instrumented element recently developed at INR-Pitesti describing technology choices as balance between fabrication complexity and risk of failure in joining processes, especially in later stages when added value increases. All joints (welded or brazed) fall into microjoining category,... and it is shown how some special provisions may influence reliability. Focus is put on brazing thin-walled Inconel sheathed thermocouples, where erosion and local loss of ductility are known issues, leading to sheath rupture. Choosing as filler the less aggressive BNi-9 helped too little. A simple but efficient technique has been developed to address this matter adequate to narrow spaces inside a nuclear fuel element, where no room is available for solutions described in literature e.g. distal preplacing of filler. The solution prevents sheath from having prolonged contact with large volume of molten filler by using locally a miniature barrier (thin stainless-steel coil or sleeve) which only allows capillary wetting, being also a perfect real-time visual indicator of brazing progress and completion. As proved in the present paper, this method along with using filler formulation with lower Carbon content (without organic binder) enhances significantly, 8 times at least, resistance to bending fatigue. A particular vacuum brazing chamber design is employed: narrow quartz tube with external induction coil and top fitting letting outside the long thermocouples attached, reducing much the chamber volume and degassing. Careful impedance match is therefore required to overcome induction power loss due to the larger coil-to-workpiece gap. Additional joining problems are discussed e.g. inherent differential expansion of long parts during induction heating which afterwards may put tension upon braze during solidification and determine delayed cracking, this being avoided through wise order of operations. Another concern is the final precision weld between instrumentation segment having attached the hard-to-handle long thermocouples bunch and nuclear segment with the heavy Uranium pellets. The result of this research is successful assembling of first Romanian prototype with joints exhibiting He leak rate bellow 1E-09 std.cc/sec and overall reliability proved during reactor irradiation testing.