Nlp Scientist - Bengaluru, India - Interactions
Description
Who is Interactions?Interactions, LLC is the world's largest independent AI company. We operate at the intersection of customer experience and AI - two of today's most innovative and dynamic industries.
Since 2004, we've helped leading companies like MetLife, Citi, and Shutterfly have millions of successful conversations, resulting in saving operational cost and increasing productivity.
Interactions' 5-year vision is to accelerate a transition from today's frustrating and uninspired customer service experiences into amazing conversational engagements Allowing customers to communicate in their own words and on their channel of choice, to accomplish tasks, all without having to go through an agent.
In doing this via our conversational AI engine, our customers benefit from dramatically improved customer experience and increased customer engagement, while also saving significant and demonstrable operational expenses.
Job Description:
Interactions is looking for a strong, talented Speech Scientist with a strong history in building Statistical Language Models, and using Machine Learning for classification.
As a member of the Speech and Language team, you will work with leading people in the industry to develop modeling techniques to advance the state of the art in spoken language understanding.
This work is high profile and will have a direct impact on the company and our industry's leading customer care Solutions.
Job ResponsibilitiesResearch methods for using deep learning in speech recognition
Use a flexible, analytical approach to design, develop, and evaluate predictive models and advanced algorithms that lead to optimal value extraction from the data.
Work with product engineers to translate prototypes into new products, services, and features and provide guidelines for large-scale implementation.
Support customer deployments of services using speech recognition and natural language processing.
Analyze reported problems or errors, identify root causes, and propose and test remedies to improve output of automated systems
Other Duties and Responsibilities:
Become an expert in our speech and natural language processing software platform, Curo.
Assist in the collection and preparation of data for training and evaluating (VB)technology.
Identify and devise strategies to accommodate language-specific features.
Develop improved methods for determining pronunciation.
Analyze supervised data from human labelers and transcribers for use in building better models.
Identify, collect, contract, or purchase training data.
Explore innovative methods to improve accuracy of speech and language systems, especially with limited data and develop creative strategies for collecting, transcribing, and formatting data.
Maintain awareness of ongoing technological research in speech recognition, language processing, and multilingual development such as low-resource Language modeling, active learning strategies and prototype the most promising advances for use in live interactive systems.
Develop supervised and unsupervised models, implement deep neural net training recipes, write development software tools, and find innovative methods to improve performance of a system that comprises speech recognition, intentclassification, and confidence scoring.
Design and run experiments to test innovations and refine testing
Qualifications:
Required:
Bachelors Degree in Electrical Engineering, Computer Sciences, or Mathematics with specialization in speech recognition, natural language processing, or machine learning.
Experience in building speech recognition and natural language processing models for large enterprise customers.
Experience training neural networks.
Possess a solid understanding of speech and language processing.
Familiar with programming and scripting languages ( Python
- Mandatory, C++
- Good to have).
Solid methodology in using utterances to build, train, and test models, with the ability to use standard speech and machine learning software and techniques.
Ability to follow processes to foster improved quality control of output models
Preferred
Masters degree in Electrical Engineering, Computer Sciences, or Mathematics with specialization in speech recognition, natural language processing, or machine learning.
Experience building multilingual acoustic and languag