前瞻AI應用開發工程師

1. AI Core & Advanced LLM Engineering

• Retrieval-Augmented Generation (RAG): Design and optimize advanced RAG pipelines; proficient in embedding model fine-tuning, strategic chunking methods, and vector database management (e.g., Qdrant, Milvus, Pinecone).
• Evaluation & Optimization: Implement re-ranking algorithms and conduct rigorous quantitative evaluations of retrieval accuracy using metrics such as Hit Rate and MRR (Mean Reciprocal Rank).
• Multimodal & OCR Processing: Architect Vision-Language Model (VLM) and OCR workflows to parse complex documents and execute structured data extraction.
• AI Agents: Design and implement autonomous AI Agent architectures, demonstrating a deep understanding of task planning, memory management, and tool/skill orchestration.

2. Software & Enterprise System Architecture

• Full-Stack Development: Build robust, scalable services using Python or C#, coupled with modern UI frameworks (Vue 3/HTML); expertise in asynchronous programming and RESTful API design.
• System & Security Architecture: Design secure, multi-tenant enterprise architectures incorporating Role-Based Access Control (RBAC), centralized logging (ELK/Loki), and optimized SQL/NoSQL database schemas.
• High-Performance Networking: Implement high-throughput, low-latency distributed communication systems utilizing TCP/IP, gRPC, or NATS.
• Middleware Integration: Leverage Redis caching strategies to optimize session management and handle high-concurrency workloads efficiently.

3. Inference Optimization & MLOps

• Inference Engines: Deploy and accelerate production-grade models using advanced inference frameworks such as vLLM, llama.cpp, Ollama, stable_diffusion.cpp or OpenVINO.
• Hardware-Aware Optimization: Apply deep knowledge of GPU/NPU architectures to evaluate the impact of compute metrics (TOPS) and KV Cache utilization
• Experience with text-to-image/video generation, Diffusion model inference, and Transformer model architecture is preferred.

※工作地點：台北(板橋)、竹南