Junior Backend Engineer (Generative AI)
Mô tả công việc
You will work directly with LLM APIs and open- source models to build AI- powered experiences that deliver real value to users, such as more relevant results, clearer answers, and consistent behavior they can trust. This requires a strong focus on relevance, stability, latency, and cost efficiency, supported by solid engineering fundamentals and system- level thinking beyond prompt- only solutions.
We are building reliable, production- grade Generative AI features that power core product experiences.
WHAT YOU’LL DO
As a Junior Software Engineer (Generative AI), you will work hands- on with real- world AI systems in production, contributing to AI- powered features across the full development lifecycle — from design and evaluation to optimization and deployment — under the guidance of senior engineers.
● Assist in optimizing trade- offs between quality, latency, and cost to enable sustainable AI usage at scale
● Build and maintain RAG pipelines using embeddings, reranking, filtering, and related techniques to ensure predictable and high- quality outputs
● Integrate LLM APIs and open- source models into scalable backend systems running in production
● Implement, track, and analyze offline and online evaluation metrics to support continuous quality improvements
● Contribute to the design and implementation of AI- powered features such as search, recommendation, and personalization, with measurable impact on relevance and user experience
● Implement guardrails, fallback mechanisms, and basic monitoring to improve system reliability and safety
Quyền lợi
Annual health checkup Health and Accident insurance 24/7 program
Supportive and stable environmentA collaborative workplace that enables engineers to focus on building long- term products.
Up to 22 full- paid days of annual leave
Cập nhật gần nhất lúc: 2026-01-22 05:15:03

NAVIGOS GROUP
Bí kíp tìm việc an toàn
Tiện ích hỗ trợ bạn
Việc làm đề xuất liên quan
Hiện tại chúng tôi chưa có việc làm đề xuất phù hợp với bạn.








