Speech Technologies Machine Learning Engineer

Key Skills

Java Programming, Python Programming, C++, PyTorch, Algorithms, Speech recognition, RNNs, CNNs

Job Description

IBM Watson Speech is an industry leader with deep expertise in automatic speech recognition and speech synthesis. We deliver that expertise to global clients with a set of publicly available Speech services that provide robust and customizable recognition and synthesis capabilities.

You will be working with a world-class team of peers from research and development to build products and services that deliver on the promise of AI by solving complex, high-value business problems. You will create and maintain machine learning models, such as RNN-T models, with a focus on speech-centric datasets and enterprise-grade performance. You will work in a highly collaborative environment, and you will use Agile methodologies to understand, develop, and iterate on solutions that meet direct customer needs. Your work will span requirements gathering & architecture, coding, testing, and deploying complex machine learning-based systems. You will employ frameworks like IBM Design Thinking to create products with user interfaces that make it easy for customers to leverage even the most advanced capabilities of a service. You will use IBM’s flagship cloud infrastructure and a development environment with the latest AI compute and networking capabilities to build services that deliver performance, quality, security, and stability.

Above all, we are looking for team members with a keen sense of ownership, resilience and drive. If you like dynamic projects and enjoy being challenged, if you can make progress when faced with ambiguous requirements and uncertainty, if you enjoy moving fast (and are not afraid of making a mistake or two on the way), then we may be the perfect team for you!

Introduction

Our mission is to revolutionize customer service by developing and deploying the latest conversational artificial intelligence (AI), natural language processing (NLP) and speech technologies in IBM’s industry-leading chatbot platform Watson Assistant.
We architect and deliver industry-leading SaaS offerings based on speech recognition and synthesis technology, with a particular focus on training and optimizing low-latency state-of-the-art machine learning models for a diverse set of customer use cases. We are actively building the next generation of our speech capabilities at IBM, and we are now significantly expanding our team!

Required Technical and Professional Expertise

• Experience with deep learning product development for speech recognition use cases -- e.g., RNNs, CNNs, sequence-to-sequence modeling for speech-to-text and/or text-to-speech
• Foundation in core Computer Science (CS) competencies including data structures, algorithms and object-oriented design
• Experience working with massive datasets involving audio data
• Experience with optimizing decoders for speech technologies
• Expertise with the PyTorch framework
• Familiarity with core concepts relevant to automatic speech recognition and synthesis, such as: speaker diarization, phonetics, speech production, signal processing

Preferred Technical and Professional Experience

• 3+ years experience as a software developer in R&D related to Speech technologies
• Experience in production-grade software development -- Java, C++, Python
• Experience deploying PyTorch models in production environments
• Knowledge of one or more non-English languages
• Familiarity with and a passion for getting your code to run in highly available cloud deployments with a focus on low-latency data processing

Apply Online at IBM AI