The past few decades have seen an increasing demand for miniaturized photonic technologies for guiding, manipulating, and analyzing light in various integrated research and commercial applications, such as telecommunications, observational astronomy, and disease diagnostics. Two application areas where photonic integration continues to lead to promising innovations are on-chip single particle biosensors and next generation spectroscopic platforms. On one hand, different integrated photonic technologies have shown great potential in point-of-care biosensing by implementing a variety of sensitive schemes in lab-on-chip platforms. On the other hand, spectral analysis has been crucial to many breakthrough innovations and discoveries, leading to significant advances in developing miniaturized spectroscopy platforms. However, several challenges remain in realizing fully integrated, compact, simple, and cost-effective photonic platforms in both these applications.In the context of pathogen sensing, liquid-core antiresonant reflecting optical waveguide (LC-ARROW)-based optofluidic biosensor platforms have demonstrated promising capabilities of detecting individual pathogens (e.g. nucleic acids, proteins, virus) from femtoliter sample volumes using different kinds of planar, single or multi-spot photonic excitation waveguides, such as single-mode, Y-split or multi-mode interference (MMI) waveguides. Additionally, the last excitation approach has also demonstrated multiplexed screening of different pathogens in the same sample using spectrally dependent multi-spot excitation. The planar excitation methods, however, face several limitations, including stable fiber-to-device coupling, spectral dependence of waveguides, and high-quality fabrication requirements. The first part of this thesis discusses the development of two different free-space, fiber-free, top-down excitation schemes. The first scheme involves excitation with a focused beam through a slit pattern milled into an opaque aluminum film covering the top surface of the LC-ARROW channel. Comparable performances for single bead fluorescence detection between this top-down multispot excitation and the planar, MMI waveguide based excitation is observed. This top-down approach also demonstrates encoded, multiplexed fluorescence detection from micro-particles with two lasers. A second top-down illumination scheme that images the spot pattern from a planar Y-split waveguide directly onto the detection device for high-fidelity fluorescence detection is also reported. This approach circumvents the need for an opaque cover and produces a further 2.7× improvement in signal-to-noise ratio compared to the first scheme. Advancements in integrated photonic spectrometers are just as crucial, offering transformative ideas in fields such as biosensing, astrophotonic integration, and environmental monitoring. The second part of this thesis introduces the idea of an integrated photonic spectrometer based on top-down imaging of an MMI waveguide combined with convolutional neural network (CNN) analysis. By capturing the wavelength-sensitive interference patterns using a top mounted camera and using CNN analysis trained on the spectra generated by known, tunable sources, this spectrometer achieves highly accurate performances in the visible and near-infrared wavelength ranges. A spectral resolution of 0.05 nm is reported in the near-infrared wavelengths, and accurate narrowband and broadband spectral reconstruction in both spectral ranges are demonstrated. The compact MMI spectrometer's capabilities are further highlighted through a 4x4 arrayed configuration on the same chip, which significantly reduces the data acquisition time and shows the scalability of this approach for simultaneous multi-target observations. A key demonstration of its applications potential is the spectral analysis of the solar spectrum, where the spectrometer successfully reconstructs the solar spectrum based on training for gas dips using a tunable laser. Next, the ongoing efforts of MMI spectrometer integration into the 3 meter Shane telescope in San Jose, California, followed by the discussion of the key challenges and potential way forward is discussed. The last major focus of this work is on improving the spectrometer’s performance in low-light conditions. Enhancing signal detection for sub-nanowatt input power levels reduces the need for expensive photodetectors, especially in applications such as astronomy or molecular spectroscopy. Selective roughening of the waveguide surface via plasma etching can enhance sensitivity and dynamic range of the MMI spectrometers by 15 dB, enabling the analysis of input test light levels as low as 300 picowatts, and also resulting in a measured scattering coefficient of 1.109 cm-1 from the etched section. It is also observed that the performance of the MMI spectrometers in such low-light applications improves with the selection of MMI pattern sections with highest pattern variations for imaging. These results highlight the potential of the MMI spectrometer for high-performance spectroscopy across disciplines.