Applied Scientist, Alexa International Tech
Core
Develop and implement algorithms and modeling techniques for Large Language Models (LLMs), ASR, TTS, and Speech-to-Speech models to advance multi-lingual applications for Alexa's international products.
Role type
Applied Scientist (Generative AI & Speech)
Builds
Production ML models for multimodal personal digital assistants
Domain
Consumer Technology, Natural Language Processing, Speech Recognition
Deliverable
production ML models
Required skills
LLM architectures, SFT, DPO, RLHF, RLAIF, ASR, TTS, Speech-to-Speech, deep learning, generative models, Python, Java, C++
Preferred skills
Top-tier conference publications, speech recognition system development, machine translation, NLP systems, professional software development
Technologies
LLMs, ASR, TTS, Speech-to-Speech, SFT, DPO, RLHF, RLAIF
Responsibilities
Analyze and model customer behavior based on large-scale data; Build and support evaluation metrics for multimodal assistants; Train and fine-tune ASR, TTS, and S2S models; Fine-tune/post-train LLMs for specific locations and languages; Set up experimentation frameworks for A/B testing; Contribute to research efforts; Collaborate on LLM evaluation frameworks; Deliver scientific solutions from research to production.
Seniority
Mid-Senior, hands-on IC