AI Engineer II for Scientific Assistant (Remote, Equity)

ArrePath; United States; Remote; USD 130,000 - 170,000; Full time

Job summary

ArrePath is seeking a second engineer to build and extend an AI assistant that supports scientists in computational chemistry, literature review, and data analysis. You will collaborate with the technical lead to take requests from design through validation and production release, ensuring robust memory, context handling, and task orchestration.

You will improve long-running task planning, tool selection, and recovery for external services, while enabling delegation to specialized agents.

Qualifications

• Practical experience building LLM agent systems, including tool use, memory, planning, delegation, and evaluation.
• Substantial experience using coding agents on a real codebase, with the judgment to review output and break work into manageable tasks.
• At least two years of professional Python or TypeScript experience.
• Clear written communication for specifications, reviews, and instructions.
• Ability to become productive quickly in a large existing codebase.

Responsibilities

• Improve how the assistant remembers information across sessions and finds relevant context.
• Help it track files, results, and other artifacts across extended research projects.
• Strengthen planning for long, multi-step scientific tasks and keep work on course.
• Improve tool selection, execution, and recovery when external services or workflows fail.
• Develop delegation to specialist agents and ways to combine their results.
• Enhance the TypeScript chat interface and deliver end-to-end changes using coding agents.

Skills

LLM agent systems

Coding agents

Python

TypeScript

Clear communication

Tools

MCP servers

Retrieval systems

Search systems

ML model training

Job description

ArrePath is seeking a second engineer to build and extend an AI assistant that supports scientists in computational chemistry, literature review, and data analysis. You will collaborate with the technical lead to take requests from design through validation and production release, ensuring robust memory, context handling, and task orchestration.

You will improve long-running task planning, tool selection, and recovery for external services, while enabling delegation to specialized agents.