About

Vishal Khadake is an experienced NLP Data Driven Applications Developer with over 7 years of relevant experience in the field. He is currently working as a Senior Data Scientist at 247.ai, where he focuses on building Machine Learning classification models using SVM and RNN for NLP data in broadcasting, satellite radios, and banking sectors. He also specializes in implementing sentiment and social models for conversational chatbots in broadcasting and hospitality sectors. Vishal started his career as a Research Intern at Speech Lab tifr, Mumbai, where he worked on a project to explore novel signal processing techniques for improving the performance of Automatic Speech Recognition (ASR) systems under noisy environments. He worked on Marathi speech data of 15000 speakers collected over 34 districts in Maharashtra and found that selective signal processing weighting under glottal closure portion can increase the recognition accuracy of ASR systems in noisy environments. As a Research Scholar at HLT-MT Group, FBK, Italy, Vishal worked on a project that focused on the importance and usage of one of the most common errors in pronunciation of foreign language by non-native students. He worked on designing a supportive network that helped an existing ASR system to provide necessary feedback to learning children, ultimately improving their pronunciation ability in the targeted language. Vishal's technical expertise includes NLP, AI, ML, research science, SVM, Python, Java, neural networks, RNN, Scikit, JavaScript, TensorFlow, Spring Boot, Spring, and Pytorch. He has also worked with NLTK and XSLT and has experience working with RDBMS and HTML. In his current role at hyr global source inc, Vishal is responsible for designing, developing, and customizing AI/NLP products and applications using advanced open-source programming languages. He has also worked as an NLP Developer at Champ Info Software, where he built an automatic medical transcription generator for medical/clinical domain data. The ASR system he built achieved 84.27% accuracy and greatly reduced manual efforts for medical transcription generation. Overall, Vishal is a highly skilled NLP Data Driven Applications Developer with a strong background in research and a proven track record of building successful ASR systems and chatbots.