Worked on performance enhancements for the pymc-devs/pytensor repository by developing Numba-accelerated sort and argsort operations for array manipulation in Python. Leveraged both NumPy and Numba to introduce JIT-compiled sorting paths, improving throughput for sorting-intensive workflows. The implementation supported axis-aware sorting and ensured compatibility with multiple sorting algorithms, while also providing warnings for unsupported sorting kinds to enhance robustness and developer feedback. Testing procedures were streamlined and coverage expanded, reducing complexity and increasing maintainability. This work reinforced a performance-oriented design, enabling faster numerical workloads in model evaluation and data processing, and demonstrated effective cross-library integration and scalability.