Shashank Mishra Email and Phone Number

Areas of work: Speech Analytics (Speech Recognition APIs, Voice Command Recognition, Voice Activity Detection, Speech-based Sentiment Analysis, Speaker Diarization), Image Processing, Explainable AIDomains: OIl-Field, BFSI, Health-careRoles and Responsibilities:ML Engineer responsibilities :Data generation / data gathering, data augmentation, data cleaning, data preprocessing, data analysis, building custom Deep Learning models, fine-tuning models, testing and deploying.Business Analyst responsibilities:Requirement gathering and understanding, analyzing business requirements and needs, designing technical and functional design documents, use case building, interacting with clients/SMEs/system architects and presenting demos.Submitted three research papers based on our work and findings from Voice Command Recognition and Speech-based Sentiment Analysis at EUSIPCO 2019, WASPAA 2019 and SIRS 2019.Work contribution:1) Benchmarked commercial and open source Speech Recognition APIs and identified the best performing API for converting huge corpora of speech data to text data.2) Contributed in developing custom Deep Learning Model (using CNN) by training on speech data generated by augmenting crowdsourced voice data contributed by Accenture employees. Model was built on Google's TensorFlow framework for an oil-field client.3) Interacted with clients to understand requirements and business problems for Speech-based Sentiment Analysis use case(finance domain client). Designed FDD and TDD, created use case diagrams and performed final testing. Also contributed in developing deep learning model for mentioned use case.4) Delivered POC for Speaker Diarization use case (nurse-patient conversation) for health care client.5) Developing and managing database using Python for Explainable AI.6) Completed Deep Learning Specialization course as a part of work curriculum and gained hands-on experience on Image dataset and Image processing application