Scale and scope of ecological research has been expanding in the past two decades. To assists in formulating and testing advanced hypothesis in this advancing field of ecology, collection of ecological data is also rapidly expanding. To keep up with growing variety, volume, velocity and, veracity of the data; new ways to manage, analyze and infer these data are being developed. These new challenges are increasingly being recognized both by ecologists and data science community. One such source of rapidly growing ecological data are remotely-triggered cameras and associated probes. Camera-trap surveys are often part of long-term monitoring programs and produce high volume of data (in the order of hundreds of thousands of images). Often such data are organized manually requiring many hours of effort. Frequently, due to effort-intensive nature, these data are under-utilized and under-analyzed. For my dissertation, I am using multi-year dataset of thousands of camera-trap imagery coming from a long term mammalian species monitoring program based in the Kameng Protected Area Complex, located within the Greater Himalayan Region (GHR) of northeast India In a collaborative project, I am using transfer learning technique to develop convolutional neural network (CNN) model for northeast Indian dataset which is upwards of 70,000 images which are unequally distributed amongst species categories. As deep neural networks (DNN) require large amount of labeled data, which is a challenge while dealing with elusive carnivores, transfer learning is flexible methodology which allows pre-trained models (e.g., ResNet, MobileNet for our purposes) to be used with our data. This process allows for higher accuracy and other metrics of model success while limiting computational time and resources. Our envisioned product of automatized image detection project is development of a generalized model for Indian subcontinent that utilizes information from species distribution models (predictor variables for species occurrence) and image data to predict species identity probabilistically.