Senior Full-Stack Developer — Agentic AI Product

American IT Systems; Jersey (AR); Hybrid; USD 110,000 - 160,000; Full time

Job summary

Agentic AI is seeking a Senior Full-Stack Developer to join our product team focused on AI-powered applications. The role emphasizes evaluating AI-generated code for correctness and alignment with house standards, using React and NodeJS.

You will operate in a hybrid New Jersey setting, with on-site days and collaboration with cross-functional teammates. Strong React/NodeJS mastery and AWS experience are essential, as is an ability to spot issues in AI-produced code and improve overall quality.

Qualifications

• Strong React and NodeJS experience
• Ability to spot mistakes in AI-generated code
• Testing/automation aptitude
• Adapt to non-linear, LLM-driven logic
• AWS infrastructure experience

Responsibilities

• Evaluate AI-generated code for correctness and alignment with house standards.
• Develop in React and NodeJS for AI-driven applications.
• Perform basic testing and automation tasks where needed.
• Collaborate on integrating AWS infrastructure.

Skills

React

NodeJS

Code review AI

Testing/Automation

AI/ML code evaluation

AWS

Tools

AWS

Job description

Senior Full-Stack Developer - Agentic AI Product

Location: Must be local to New Jersey - hybrid, 3 days/week on-site
Interview: Must be local for on-site interview
6+ Months

Required Skills:

• Strong React and NodeJS experience (explicitly React + nodeJS - not Java).


• Must be able to independently spot mistakes in AI-generated code - this is described as the single most important skill for the role, more important than raw code-writing speed.


• Some testing/automation aptitude, developers are expected to be able to do some testing themselves, not just hand off to a dedicated QA function.


• Ability to adapt to non-linear, LLM-driven application logic rather than traditional deterministic code paths.


• AWS infrastructure experience.

This role is focused on React, NodeJS, and critically evaluating AI-generated code for correctness, structure, and adherence to team standards and patterns. The code generation itself is described as \"pretty good\" and functional - the real skill gap is knowing whether it is right according to house standards, not whether it runs.