With the advent of machine learning, statistical learning and modelling becoming more prevalent in today’s world, we can estimate values for an unknown data given a trained model i.e. finding value of data(D) given parameter (Θ) .Now what about the reverse? What if we want to find optimal parameter value, given a particular set of dataset. Why is this even required? The very simple answer to this can be found in every modelling and computation textbook. We cannot necessarily measure the other parameters. Let us take an example: There are several parameters while defining a neuron whose measurement is difficult to make, for example the axial resistance, the capacitance of the soma etc. Thus to achieve this process we make use of 2 tools, the first is the neuron library which is used for designing your neuron model and setting your parameters ,secondly is a statistical measurement tool used for the parameter estimation .This is done by checking for the voltage value generated by the neuron model and what might the parameter values be for that particular value. The reason to utilize Bayesian inference for parameter estimation is due to the specification of the prior data ,which will help us to increase the accuracy of our model which predicts approximate parameter values. Now the user could specify the prior data on the basis of any previous study or research material. Neuroptimus is an open source tool created for the same purpose. It is an interactive tool where the user can drop a dataset file. Train it and use it for parameter estimation . I have been tasked to utilize the prior ,the noise (for the calculation of likelihood),and the evidence to find the posterior distribution of the parameter I want to estimate. This is the main gist of the project, but there are various hurdles such as what if it is difficult to find the likelihood ,type of noise, computational costs for multiple parameters etc. These are some other things I believe I will be working on.