Transition metal dichalcogenides (TMDs) are semiconductor Van der Waals materials that can be obtained as stable monolayers (ML), characterized by bandgap values from visible to near infrared range. With decreasing thickness from bulk to MLs, TMDs show a transition from indirect to direct bandgap, which combined to strong exciton binding energies makes them propitious for optoelectronic applications. Their atomically flat, dangling bond free surfaces and weak interlayer van der Waals bonding enable stable exfoliation or growth of uniform MLs, which combine high oscillator strength, short radiative lifetimes, and spin-valley dependent band structures. Strain engineering further expands the functional palette of TMDs, modulating bandgaps, enhancing carrier mobility, and narrowing photoluminescence linewidths. At the same time, spatially localized strain gradients funnel excitons into nanoscale potential wells, boosting emission efficiency and enabling deterministic single-photon sources. However, most demonstrations rely on passive substrates or optically pumped mechanisms, precluding scalability and electrical control. Electrically injected TMD devices would enable reduced power consumption, continuous operation, and lithographically defined injection regions, but require robust control of charge transfer, interface band alignment, and defect formation. Optical and electronic behavior of 2D-TMDs is strongly influenced by substrate chemistry and morphology. Most studies integrate TMDs on amorphous oxides, but combination with epitaxially grown III-V materials offers a dual opportunity: precise strain patterning via three-dimensional nanostructures, and engineered heterojunctions for efficient charge injection and extraction. In this sense, selective area epitaxy (SAE) allows scalable growth of vertical nanowires, horizontal nanomembranes, and interconnected networks, with precisely controlled position and facet orientation. Engineering shape and size of III-V nanostructures requires deep understanding of the formation and epitaxial growth mechanisms. A quantitative comprehension of SAE growth allows control of the nanostructure morphology, thus determining strain distribution and interface confinement. Such knowledge on 2D-TMD/III-V heterostructures would prompt the design of conformal, strain engineered junctions for direct electrical pumping of tunable localized emitters. Adequate characterization is fundamental to address the heterojunction properties. Considering limitations of optical probing and complexity of contact deposition on 2D-TMDs, electron beam induced current (EBIC) stands out as an optimal way to evaluate the heterojunction electrical behaviour. Without damaging active regions of interest and with nanometric resolution, EBIC allows to acquire knowledge of charge dynamics and electric fields at nanoscale interfaces. In this thesis, 2D-TMDs are integrated onto both planar and nanostructured III-V substrates to study the interaction of interface nature and electronic coupling. I present scalable templates of disparate morphologies for nanoscale charge injection, and I describe a reliable, not invasive characterization method to access electric fields at the nanoscale. By correlating nanostructure geometry and carrier distribution, I show controlled carrier injection into localized regions. These results lay the foundation for high performance, scalable 2D-TMD/III-V heterostructures in optoelectronic and quantum devices.