Java Web3 Jobs in Taiwan

This role sits at the intersection of quantitative finance, machine learning, and natural-language understanding. You’ll leverage NLU techniques—like sentiment and intent analysis—to mine news and social media for trading signals, build and mathematically refine predictive models, and rigorously backtest and optimize your strategies to drive data-driven investment decisions.

Responsibilities:

• Responsible for the architecture design and core module development of LLM-based AI Agent systems, including task planning, memory management, tool invocation, and workflow execution;
• Lead the development of a general-purpose Agent platform capable of multi-task coordination and multi-agent collaboration, enhancing the usability and stability of large models in complex business scenarios;
• Leverage key technologies such as LLMs, RAG, vector search, and function calling to build intelligent systems with contextual understanding and knowledge injection capabilities;
• Develop evaluation systems and automated testing tools for Agent behavior to continuously optimize task completion rate, response quality, and reliability;
• Drive the application of AI Agents in vertical scenarios such as search, trading, deep analysis, and automated data engineering to unlock the full value of LLMs.

Requirements:

• Bachelor’s or Master’s degree in Computer Science, Mathematics, Statistics, Financial Engineering or a related discipline
• Solid understanding of LLM principles (e.g., Transformers) and their application in Agent systems, with end-to-end experience in building and launching intelligent agents;
• Proficiency in at least two mainstream open-source Agent frameworks such as LangChain, AutoGen, LangGraph, ADK, LlamaIndex, Dify, CrewAI, or CAMEL, with a strong grasp of their internal architectures;
• Familiar with RAG (Retrieval-Augmented Generation) and capable of designing and optimizing knowledge retrieval systems; experience with vector databases like Milvus or FAISS is required;
• Experience implementing key Agent modules including Tool Use, Task Planning, Memory Management, and understanding their integration in real-world applications;
• Strong software engineering background, proficient in Python/Java, and familiar with client-server architectures and MCP-like communication protocols.

By submitting a job application, you confirm that you have read and agree to our Candidate Privacy Notice.