Fine-tune LLM mã nguồn mở bằng LoRA, QLoRA, PEFT, SFT
Xây dựng giải pháp OCR, Document Layout Analysis, Information Extraction
Thiết kế và triển khai hệ thống AI Agent, Multi-Agent cho quy trình phức tạp
Tối ưu pipeline huấn luyện, đánh giá và inference mô hình
Có thì tốt
AWS, Azure, MLOps, distributed training
Multimodal AI (VLM) và fine-tuning Document AI
Thiết kế Agentic AI production-grade (RAG, planning, memory, multi-agent)
Tóm tắt bằng AI từ JD gốc, đọc JD gốc trước khi ứng tuyển.
Mô tả công việc
We are looking for a skilled AI Engineer with a strong foundation in Machine Learning, Deep Learning, and NLP. The ideal candidate has hands-on experience in LLM fine-tuning, OCR & Document Understanding, and AI Agent/Multi-Agent systems, with the ability to develop, optimize, and deploy AI models in production. This role focuses on Core AI/ML Engineering, LLM Fine-tuning, and Agentic AI. Candidates are expected to have hands-on model training and fine-tuning experience, rather than solely LLM API integration, prompt engineering, or RAG chatbot development.
Research, develop, train, fine-tune, and deploy AI/ML models.
Fine-tune open-source LLMs using LoRA, QLoRA, PEFT, and SFT.
Develop OCR, Document Understanding, and Information Extraction solutions.
Design and implement AI Agent and Multi-Agent systems for complex workflows.
Build and optimize model training, evaluation, and inference pipelines.
Research and implement advanced AI techniques from academic papers.
Yêu cầu công việc
Education: Bachelor's degree/Master’s degree in Data Science, Artificial Intelligence, Computer Science, or a related field.
Experience: Minimum 1–2 years of hands-on experience in AI/ML (industry experience)
Programming: Python, TypeScript/JavaScript, SQL.
ML Frameworks & Libraries: PyTorch, TensorFlow, NumPy, Pandas.
AI/Deep Learning: Strong understanding of ML, DL, NLP, and Transformer architectures.
LLM Fine-tuning: Hands-on experience with LoRA, QLoRA, PEFT, and SFT on models such as Qwen, Llama, and Mistral.
AI Agents & Multi-Agent Systems: LangGraph, LangChain, CrewAI, or equivalent; tool calling, agent orchestration, multi-agent collaboration.
Document AI: OCR, Document Layout Analysis, and Information Extraction.
Model Optimization: GPU/CUDA, quantization, hyperparameter tuning, model evaluation.
Deployment: FastAPI, RESTful APIs, vLLM, or equivalent frameworks.
Experience with AWS, Microsoft Azure, MLOps, and distributed training.
Experience with Multimodal AI (VLM) and Document AI fine-tuning.
Experience designing production-grade Agentic AI systems, including RAG, planning, memory, and multi-agent orchestration.
Publications at CORE Rank A/A* conferences (e.g., ACL, EMNLP, NAACL) are a strong advantage; conference ranks should be checked against the relevant ranking edition.
Ability to read, reproduce, and implement methods from AI research papers.
