Applied Scientist

Summary

I have worked at Impact Tech LTD, a startup in the field of conversational AI, as machine learning research engineer. During this time, I proposed, evaluated and developed machine learning models for Automatic Speech Recognition (ASR) which outperformed Amazon AWS and Microsoft Azure ASRs in terms of word error rate. As a result, this led to my employment at Amazon Alexa perceptual technologies. Under the Alexa Perceptual Technologies organization at Amazon, I completed a project that has purpose to improve Speech Emotion Recognition. Recent challenges in the literature suggest that current methods in Emotion Recognition struggle from the fact that ML models have over-reliance in specific modalities of data over others (e.g. Text over Speech). As a result, some ML systems do not perform adequately well (e.g. "I love you" in an angry voice is detected as "Happy"). My contribution is a novel method towards disentanglement of data modalities. I proposed, developed and evaluated models with this method that outperformed all current state-of-the-art methods in Language-agnostic Speech Emotion Recognition by a significant amount. The method is widely applicable to other areas and not limited to Speech. The project's scope is improving Alexa's services in the Emotion Recognition domain. This work was submitted to ICASSP.

Expectations

Looking for a role in applied machine learning, quantitative research, statistical modeling, deep learning.

Employment Preferences

Expected Base Salary

**0,000 USD

Academic Degree

Experience

Skills

• Machine Learning
• Artificial Intelligence
• Applied Research
• Applied Science
• Research Engineer
• Quantitative Researcher
• Big Data
• Python
• C++
• NLP
• Automatic Speech Recognition
• ASR
• Speech Emotion Recognition
• Signal Processing
• Reinforcement Learning
• Supervised Learning
• Unsupervised Learning
• Speech Processing
• Computer Vision
• Multimodal Machine Learning
• Self-supervised Learning