4 months ago
Sŏngnam, Korea, SouthSenior
Responsibilities
- Perform LLM fine-tuning and post-training for Navigation Agents, including instruction tuning, supervised fine-tuning, and preference optimization.
- Develop agent models for intent and information extraction, navigation function execution, tool and API selection, execution planning, and context-aware responses and actions.
- Develop classification, prediction, ranking, and other traditional machine learning models for agent components.
- Build and improve training and evaluation datasets and data quality.
- Design agent performance evaluation systems and key metrics, analyze failure cases, and improve model quality.
- Optimize model accuracy, stability, and response speed in production services.
- Collaborate with Navigation, Backend, Data, and Product teams to apply and continuously improve models in products.
Requirements
- Experience in research or development related to machine learning or deep learning.
- Hands-on experience with LLM fine-tuning or post-training.
- Experience designing, training, evaluating, and optimizing models using PyTorch or other deep learning frameworks.
- Experience developing and improving classification, prediction, ranking, or recommendation models.
- Ability to design and build training data and resolve model issues through quantitative and qualitative evaluation.
- Ability to learn new technologies and problems quickly and collaborate effectively with colleagues across disciplines.
- Preferred: experience developing navigation, map, or mobility services.
- Preferred: understanding of navigation domains such as route search, POI search, place recommendation, ETA prediction, and traffic information.
- Preferred: experience developing LLM-based agents, tool-use models, or function-calling models.
- Preferred: experience applying and operating search, recommendation, ranking, or personalization models in production services.
- Preferred: experience developing models using large-scale user logs or location and movement data.
- Preferred: experience building and operating distributed GPU training and large-scale training pipelines.
Benefits
- The interview process consists of document screening, a first video interview, a second in-person or video interview, and offer discussion and onboarding.
- A three-month probationary period may apply.
- Veterans and employment-protected individuals are given preference according to applicable laws, and registered persons with disabilities are also given preference.
Tech Stack
Categories
About 42dot
42dot builds software-defined vehicle and autonomous mobility platforms for automakers and mobility service providers. Its products include AI-driven vehicle software, data-based update services, and core platform components such as API gateways and identity/access management to connect services securely. The privately held company is headquartered in Seongnam-si, Gyeonggi-do, and focuses on transportation and automotive markets.