ML and AI Engineer

關於此職位

The AI analyst is the analytical heart of the Drusus platform. The work of this role is principally the work of making a frontier language model behave with the precision and discipline that financial analytical work demands: through system-prompt engineering, evaluation infrastructure that catches failure modes before they reach users, and the verification layer that constrains the model to the verified data context.

職責

• Own the system-prompt architecture of the AI analyst chat and the Drusus Daily editorial pipeline.
• Build and maintain the evaluation infrastructure that monitors model output for accuracy, regulatory compliance, and hallucination.
• Lead the integration with Anthropic Claude (the Institution tier) and the secondary models used at lower tiers.
• Stay current with the published research on language-model safety and evaluation; bring relevant techniques into our practice.
• Collaborate with the editorial team on the human-AI editorial workflow.

我們所尋求者

• Substantial experience working with large language models at the application layer, whether through Anthropic, OpenAI, or comparable APIs.
• Demonstrated ability to design system prompts that elicit reliable behaviour from frontier models.
• Familiarity with the published literature on language-model evaluation (LLM-as-judge, ground-truth datasets, adversarial testing).
• Strong general programming skills, preferably in TypeScript and Python.

額外有助但非必要

• Prior work in a financial or regulated domain.
• Experience with retrieval-augmented generation and grounded LLM systems.
• Familiarity with constitutional AI and similar alignment approaches.