AI Amplify Researcher

The Metabit Amplify Team is an elite squad dedicated to building AI-native engineering infrastructure. We design and implement end-to-end AI coding agent systems that multiply every engineer's productivity by 3-5x.

We don't write trading strategies—we build tools and systems that multiply strategy engineers' efficiency. We are also firm believers and practitioners of open source:

• Embracing Open Source: We stand on the shoulders of giants, actively adopting and integrating cutting-edge open source technologies to stay in sync with the global AI community.


• Contributing Back: We don't believe in building in isolation. As we refine our systems, we will progressively open source this secure, reliable, and efficient AI-native engineering infrastructure.

Responsibility

• Researching LLM-based coding agents, including key capabilities such as planning, tool use, memory, and long-horizon reasoning;


• Exploring how AI understands and manipulates real-world large-scale codebases and development toolchains;


• Designing new agent architectures and algorithms to push the boundaries of AI-automated software development capabilities;


• Building evaluation systems and benchmarks to measure the capabilities of coding agents in tasks such as bug fixing, feature implementation, code migration, and test generation;


• Rapidly prototyping research ideas and validating them in real-world engineering environments;


• Analyzing the capabilities and limitations of cutting-edge models (Claude, GPT, Codex, etc.);


• Tracking and reproducing the latest research results (from OpenAI, Anthropic, DeepMind, and open-source communities);


• Collaborating with engineering teams to translate research results into practical systems;


• Participating in the broader research community through open-source projects, technical reports, or papers.

We expect you to:

• A solid foundation in computer science (algorithms, systems, machine learning, etc.);


• Strong interest in LLM, Agent Systems, and AI coding;


• Experience in LLM or AI agent-related projects;


• Be able to design experiments and evaluate the capabilities of AI systems;


• Excellent programming skills (Python preferred);


• Be able to quickly understand papers and new technologies and implement prototypes;


• Highly self-motivated and enjoy exploring open-ended questions.

Bonus points:

• Research experience in LLM, code modeling, program synthesis, automated debugging, etc.;


• Published papers or shared technical information related to ML/Systems/Software Engineering;


• Familiar with large codebases or developer infrastructure;


• Participated in open-source AI projects;


• Experience with AI coding tools (Copilot/Claude Code/SWE-agent/Devin, etc.).