New opportunity

AI Software Engineer, Full-Stack (NY)

Loop · New York, United States

About this role

Loop is seeking an AI full-stack engineer to build customer-facing solutions for logistics billing and payments. The role combines shipping products that use AI with building the systems that enable AI agents and non-technical users to complete technical work reliably. The engineer will work closely with founders, customers, and cross-functional teams, with guidance and feedback from the engineering team.

Responsibilities

Partner with product, design, and sales to understand requirements and deliver solutions to customer problems.
Build an AI harness comprising data pipelines, schemas, validation, and feedback loops so agents can work reliably at scale. Define outcomes, data, and success criteria, then enable agents to write and validate solutions against real data.
Write extensible code across frontend, backend, infrastructure, and the AI stack. Participate in code reviews, lead technical design discussions and documentation, and share engineering practices.
Use tools such as Cursor, Codex, and Claude to accelerate development and encourage their adoption across the organization.

Qualifications and technology:

Requires two years of software engineering experience; experience with TypeScript and tools such as GraphQL, Relay, React, and Node.js; and experience shipping products end-to-end at an early-stage company.
Interest in applying AI to real customer problems is required. Prior AI or machine learning experience is not required; the engineer will learn Loop’s AI stack.
The stated stack includes AWS, CDK, Fargate ECS, Prisma, NestJS, PostgreSQL, Kafka, Redis, Elasticsearch, Ant Design, Vite, LLMs, prompt engineering, agentic workflows, and Loop’s internal AI platform.

Location and compensation:

Work in the New York City office at least four days a week. Base pay is $145,000–$205,000 annually, with actual pay depending on job-related factors. The role may be eligible for equity and includes health, dental, and vision insurance, 401(k) match, paid time off, paid holidays, and parental leave. Employment visa sponsorship may be available; green card sponsorship depends on business needs and eligibility.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

TypeScriptGraphQLRelayReactNode.jsAWSCDKFargateECSPrismaNestJSPostgreSQLKafkaRedisElasticsearchAnt DesignViteLLMsPrompt engineeringAgentic workflowsCursorCodexClaudeData pipelinesSchema designValidationCode reviewFull-stack development

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Solve customer billing and payment problems
The role works with founders and customers on logistics billing and payment pain points and ships customer-facing features. Cross-functional work with product, design, and sales is part of translating those needs into solutions.
Build a reliable AI harness
Loop calls for data pipelines, schemas, validation, and feedback loops that let agents work reliably at scale. The role starts with outcomes, data, and success criteria, then validates agent-produced solutions against real data.
Ship extensible work across the stack
The engineer writes code across frontend, backend, infrastructure, and the AI stack. The posting also calls for end-to-end product shipping, code reviews, technical design discussions, and documentation.
Use AI tools to improve development
Loop expects the engineer to use tools such as Cursor, Codex, and Claude to accelerate development and encourage their adoption across the organization. Prior AI or machine-learning experience is not required.

Likely challenges to think through

Balancing agent autonomy with reliable results
Likely preparation challenge, not a confirmed problem: letting agents write solutions creates a trade-off between flexibility and predictable outcomes. Prepare to explain how you would define success criteria, validate against real data, and use feedback to catch failures.
Choosing reusable solutions over one-off fixes
Likely preparation challenge: the posting favors solving a category of customer problems, which may take more design work than shipping a single feature. Prepare to discuss how you would distinguish shared patterns from customer-specific needs while still delivering useful work.
Maintaining quality while working across boundaries
Likely preparation challenge: spanning UI, backend, infrastructure, and AI work can make changes harder to coordinate and review. Prepare to describe how you would keep interfaces clear, document design decisions, and use code review without slowing delivery unnecessarily.

Practice your answers

Technical practice

How would you design an end-to-end feature for a logistics billing or payment problem?

What to coverThe role involves shipping customer-facing features across frontend, backend, infrastructure, and AI components. Preparation suggestion: explain how you would clarify the outcome, define component boundaries, and test the feature with representative data.

How would you turn fragmented or unstructured logistics data into a reliable input for an AI agent?

What to coverLoop describes a logistics data platform and calls for data pipelines in its AI harness. Preparation suggestion: discuss ingestion, data quality checks, traceability, and how you would handle incomplete or conflicting inputs.

What schemas and validation would you put around agent-generated work?

What to coverThe posting explicitly names schemas and validation as parts of the AI harness. Preparation suggestion: distinguish input constraints from output checks, and explain when an invalid result should be rejected or reviewed.

How would you define success criteria for an agent before specifying its implementation?

What to coverThe role emphasizes starting from outcomes and validating agent-written solutions against real data. Preparation suggestion: describe measurable acceptance criteria, representative test cases, and the limits of an evaluation set.

How would you troubleshoot an agent that produces inconsistent results at scale?

What to coverThe posting calls for feedback loops that help agents work reliably at scale. Preparation suggestion: explain how you would isolate data, prompt, and validation failures, then use observed failures to improve the harness.

How would you structure a React feature that reads and updates data through GraphQL and Relay?

What to coverReact, GraphQL, and Relay are named in the role's frontend stack. Preparation suggestion: cover data requirements, loading and error states, and trade-offs between reusable components and feature-specific code.

How would you design a TypeScript backend API backed by PostgreSQL?

What to coverThe listed backend stack includes TypeScript, Node.js, NestJS, Prisma, PostgreSQL, and GraphQL. Preparation suggestion: discuss API contracts, data modeling, transaction boundaries, and validation without assuming a particular existing architecture.

When might you use Kafka, Redis, or Elasticsearch in a full-stack system?

What to coverAll three appear in Loop's backend stack, but the posting does not specify their use in this role's features. Preparation suggestion: compare event processing, fast-access data, and search needs, including when simpler approaches would suffice.

How would you deploy and operate an extensible service on the listed infrastructure?

What to coverThe role spans infrastructure, and Loop lists AWS, CDK, and Fargate ECS. Preparation suggestion: discuss repeatable deployment, configuration, monitoring, and how you would diagnose a failed release.

How would you use an AI coding tool without compromising code quality?

What to coverThe posting names Cursor, Codex, and Claude as tools for accelerating development. Preparation suggestion: explain where such a tool helps, how you would verify its output, and what belongs in tests and code review.

Behavioural practice

Tell me about a time you reconciled different stakeholder needs to ship a feature.

What to coverThe role works with customers and with product, design, and sales to understand requirements. Preparation suggestion: use a real situation, explain how you resolved competing needs, and state the result.

Describe a time you took an ambiguous problem through to a shipped solution.

What to coverThe posting values end-to-end shipping in an early-stage setting and solving a category of problems rather than one isolated request. Preparation suggestion: describe the ambiguity, the scope decisions you made, and the observed result.

Tell me about a technical design discussion you helped bring to a decision.

What to coverThe role includes leading design discussions and documenting decisions. Preparation suggestion: give the situation, the alternatives and trade-offs you communicated, and the resulting decision or outcome.

Describe a time code-review feedback changed your approach or improved a team's practice.

What to coverThe posting calls for code reviews, deliberate feedback, and sharing engineering practices. Preparation suggestion: explain the feedback exchange, what you did differently, and its concrete result.

Tell me about a time you introduced a development tool or practice to others.

What to coverThe role includes encouraging adoption of AI tools to improve productivity. Preparation suggestion: use a real example, describe how you evaluated and shared the practice, and distinguish observed benefits from expectations.

Sharpen your knowledge

Topics to prepare

Outcome-led AI harness design
The role calls for data pipelines, schemas, validation, and feedback loops that let agents work reliably at scale. As a suggested exercise, sketch an agent workflow starting with the desired outcome, real-data inputs, success criteria, and checks for incorrect results.
Extensible full-stack architecture
The posting emphasizes code across frontend, backend, infrastructure, and AI components. As a suggested exercise, diagram how a customer-facing feature crosses those boundaries and identify where clear interfaces would make it easier to extend.
GraphQL and Relay data flow
Experience with GraphQL, Relay, React, and Node.js is listed. As a suggested exercise, trace a customer action from a React component through a Relay query or mutation to a Node.js service, including loading and error states.
Customer problems in logistics billing and payments
Loop says this role works with founders and customers on logistics billing and payment pain points. As a suggested exercise, turn a hypothetical customer complaint into an outcome, clarifying questions, and acceptance criteria; do not assume the example reflects Loop’s actual workflows.
End-to-end delivery and technical decisions
The posting calls for early-stage experience shipping products end-to-end, plus leading design discussions and documentation. As a suggested exercise, outline a feature’s scope, trade-offs, rollout considerations, and the decisions worth recording for reviewers.

Skills to revise

Build a TypeScript full-stack feature
Practice implementing a small React interface backed by a TypeScript/Node.js GraphQL operation, using Relay for data access. This directly revises technologies named in the qualifications; the feature itself is a suggested exercise.
Implement validation and feedback
Build a small pipeline that accepts sample records, validates them against a schema, records failures, and uses those failures to improve the checks. This practices the harness components explicitly assigned to the role; the sample records can be fictional.
Review code across component boundaries
Practice reviewing a frontend-to-backend change for correctness, extensibility, and clear interfaces, then write a short design note explaining a trade-off. Code reviews, cross-boundary work, and design documentation are stated responsibilities.
Verify AI-assisted development
As a suggested exercise, use one named tool—Cursor, Codex, or Claude—to draft a small change, then inspect, test, and revise its output. The posting expects adoption of these tools to accelerate development, not unverified acceptance of generated code.

Your preparation checklist

Prepare an end-to-end product walkthrough
Choose a product you actually shipped and outline the customer problem, your work across the stack, tradeoffs, launch, and results. Loop specifically seeks early-stage experience shipping products end-to-end.
Sketch a logistics billing design
As a practice exercise, diagram a hypothetical customer-facing billing or payment workflow, including a React interface, GraphQL/API boundary, backend, and data model. Explain where you would clarify requirements before choosing an implementation.
Design a reliable AI harness
Prepare a one-page design for a hypothetical agent task: define its outcome and success criteria, data pipeline, schemas, validation against real data, and feedback loop. These are the harness components named in the role.
Gather evidence for the stated technologies
Make a short list of work you can discuss using TypeScript, React, GraphQL, Relay, or Node.js, where applicable. Review relevant choices from Loop’s broader listed stack without implying experience you do not have.
Rehearse a technical design and code review discussion
Practice explaining an extensibility tradeoff across frontend, backend, or infrastructure, then identify tests and review feedback you would seek. The role includes design discussions, documentation, and code reviews.
Prepare an AI-assisted development example
If you have used Cursor, Codex, or Claude, outline a specific task and how you checked the output. Otherwise, prepare a proposed workflow; Loop names these tools but does not require prior AI or machine-learning experience.
Write questions for the interviewers
Ask how product, design, sales, and customers define successful solutions; how the team measures agent reliability; and what guidance engineers receive while learning Loop’s AI stack.

Availability last checked: Nov 4, 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in United States. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…