As a member of one of our Technology teams, you will contribute to building solutions that use natural language processing, cognitive computing, and artificial intelligence applications or the frameworks and infrastructure that support them.
The Principal Inventive Scientist is responsible for conducting original research by developing novel algorithms, implementing them as software tools, and building models for improving accuracy of conversational AI technologies. They are expected to participate and publish their research at technical conferences.
Job Responsibilities
-
Using data from Interactions’ vast human-annotated speech and text databases develop high-performance models and algorithms, including acoustic, language, confidence, and intent classification models to support integration of the results into commercial products.
-
Responsible for making improvements to speech and language software and modeling methods and to invent new methods to improve performance.
-
Work with and advance technology in automatic speech recognition, deep neural networks, machine learning, acoustic modeling, language modeling, machine learning, and modeling with unsupervised and unstructured data.
-
Actively participate with team members to develop innovations that push the edge of science forward.
-
Implement new algorithms, format and process large text/audio databases, run experiments to measure performance, and support implementation of successful results into products
Qualifications
Required
-
PhD or MS with industry experience in computer science, engineering, linguistics, or a related field
-
At least five years of direct experience
-
A solid understanding of speech and language processing
-
A strong publication record of research through papers and presentations
-
Strong software background in Python, C/C++, and Linux shell
-
Experience in applying deep neural networks (CNNs, RNNs, LSTMs, etc.) or other deep learning methods to speech recognition or natural language processing.
Preferred
-
Experience with speech and language or other machine learning software such as Kaldi, Pytorch, TensorFlow, or scikit-learn is a plus
-
Familiarity with a non-English language
-
Strong record of collaboration and team-work