Deep Learning

Responsibilities

• Develop and refine models for common wake-up words such as "Hey Hyundai," "Hey Kia," "Hey Genesis," and future wake-up words planned for development and commercialization.
• Collect or curate diverse intonation data from native speakers and use it to develop a localized wake-up word detector.
• Gather hard negative words commonly used in the local region and integrate them into the modeling process.
• Quickly implement the latest wake-up word technologies into real-world services and collaboratively develop technology and code for application in other locales.
• Actively communicate and troubleshoot with the Korean headquarters and Hyundai Motor Company-related entities to ensure accurate support for North American wake-up words.

Qualifications

• Over 5 years of experience in research, development, and commercialization of wake-up words.
• Experience in knowledge distillation and inference optimization for small models.
• Proficiency in Pytorch, as well as libtorch, onnx, and other on-device libraries, and familiarity with writing C++ code.
• Korean language proficiency in professional research or work

Preferred

• Experience in the automotive/mobility industry, with a focus on voice technology or related areas.
• Familiarity with continuous integration and deployment pipelines.
• Strong problem-solving skills and ability to work in a global collaborative working environment.

Recruiting Process

• Application Screening - Coding Test - First Interview - Second Interview - Offer Negotiation - Final Acceptance
• The recruitment process may differ based on the specific role and may be subject to changes depending on schedules and circumstances.
• Applicants will be notified of the application schedule and results individually via the email address provided in their application.