Worked on enhancing the LogitNormal distribution within the pymc-devs/pymc repository by implementing automatic log probability calculations and adding inverse CDF (icdf) support. This development streamlined inference workflows by enabling percentile-based queries and reducing the need for manual derivations in statistical modeling. The approach focused on improving numerical stability and reliability in posterior estimation, making the LogitNormal distribution more expressive for real-world applications. Collaborated on the implementation using Python, leveraging expertise in probability distributions and statistical modeling. The work delivered a robust feature that supports logprob-based inference and downstream sampling, contributing to more flexible and maintainable probabilistic models.