Staff AI Applications Engineer

CarParts.com•Long Beach, CA

20h

About The Position

CarParts.com is seeking a Staff AI Applications Engineer to join their team. This role is at the intersection of full-stack engineering and applied AI, focusing on building intelligent agents, integrating LLMs into their platform, and developing MCP (Model Context Protocol) servers for automation. The company is investing heavily in AI-powered engineering and this role will contribute to shipping AI features that millions of customers interact with. The ideal candidate is a builder who prefers creating reusable abstractions and scalable systems, a troubleshooter who can quickly pinpoint issues, and an AI enthusiast who applies rigor to token budgets and prompt design.

Requirements

5+ years of experience in full-stack web application development using Node.js, JavaScript, TypeScript, and modern frameworks.
Extensive experience building scalable applications and microservices using React, Next.js, Node.js, Express, HTML, and CSS.
Hands-on TypeScript across frontend and backend systems.
Strong knowledge of RESTful API design and OpenAPI specifications.
Experience designing and integrating APIs, including REST and modern data-fetching patterns.
Extensive experience with MySQL, MongoDB, PostgreSQL, and Redis, with solid understanding of data modeling trade-offs.
Familiarity with micro-frontend architecture and module federation.
Strong experience building performant React applications using hooks and state management (Redux or equivalent).
Experience with cloud-native development using Docker and containerized environments.
Experience with CDNs, caching strategies, performance optimization, and security considerations.
Strong knowledge of JavaScript build tools (Webpack, Vite, or modern bundlers).
Proficiency with Chrome DevTools and frontend performance profiling.
Experience with SPA, PWA, responsive design, and MPA architectures.
Strong foundation in data structures, algorithms, and database design.
Proven experience in software architecture, design patterns, and engineering best practices.
1+ year hands-on experience integrating LLM APIs (OpenAI, Anthropic, or equivalent) into production or near-production systems.
Demonstrated ability to build or extend AI agents that use tool-calling, function execution, and structured output.
Solid understanding of agentic concepts and design patterns: ReAct (Reason + Act) loops, chain-of-thought planning, and step-by-step task decomposition.
Tool orchestration — selecting, invoking, and chaining external tools based on model reasoning.
Memory architectures: conversation context, scratchpads, vector-backed long-term recall.
Self-correction and reflection — agents that detect errors in their own output and retry.
Human-in-the-loop checkpoints, confidence thresholds, and graceful fallback to manual workflows.
Multi-agent coordination — delegating subtasks across specialized agents and merging results.
Acquaintance or hands-on experience developing agents: Built, extended, or shipped at least one agent (production, internal tool, or well-scoped prototype) that performs multi-step autonomous tasks.
Familiar with agent frameworks such as LangChain, LangGraph, CrewAI, Autogen, Claude Agent SDK, or custom orchestration loops.
Comfortable designing agent tool schemas, managing agent state, and debugging non-deterministic agent behavior.
Experience designing or contributing to MCP servers or similar context-orchestration layers.
Proven approach to token budget management: prompt optimization, caching strategies, and cost monitoring.
Comfortable using AI coding assistants (GitHub Copilot, Claude Code, Claude Cowork, Cursor) daily to accelerate development.
Able to write effective prompts for code generation, refactoring, test creation, and documentation.
Understands foundational LLM concepts: tokens, temperature, context windows, embeddings, and RAG.
Can evaluate AI-generated code for correctness, security, and performance — not just accept output blindly.

Nice To Haves

Experience with public cloud services (AWS, Azure, GCP).
Experience with ecommerce/retail purchase journeys.
Experience migrating legacy applications to modern stacks.
Familiarity with vector databases (Pinecone, Weaviate, pgvector) and RAG pipelines.
Experience with agent frameworks (LangChain, LangGraph, CrewAI) or custom orchestration loops.
Experience fine-tuning or distilling models for domain-specific tasks.
Contributions to open-source AI tooling or the MCP ecosystem.
Experience with GraphQL.

Responsibilities

Design, develop, and own full-stack features across React/Next.js frontends and Node.js/Express microservices.
Build AI-powered product experiences: intelligent search, personalized recommendations, automated content generation, and conversational commerce flows.
Design and develop agentic systems — agents that plan, reason over multiple steps, select and call tools, handle errors autonomously, and escalate to humans when confidence is low.
Implement agentic patterns: ReAct loops, chain-of-thought planning, reflection/self-critique, memory (short-term context and long-term retrieval), and multi-agent coordination.
Develop and maintain MCP servers that expose ecommerce domain tools (catalog, pricing, inventory, order) to LLM-powered clients.
Integrate LLM APIs into production paths with proper error handling, fallback strategies, and cost guardrails.
Write prompts, evaluation harnesses, and monitoring for AI features — treat them as code, version them, review them.
Optimize frontend performance: Core Web Vitals, page load time, time-to-first-byte.
Design APIs (REST, OpenAPI) that are clean, well-documented, and backward-compatible.
Build shared tooling: CLI utilities, code generators, reusable component libraries, and internal developer tools powered by AI.
Improve CI/CD pipelines, containerized builds, and deployment workflows.
Participate in architecture decisions, code reviews, and technical design documents.
Translate business requirements into technical designs with product and design stakeholders.
Mentor engineers on AI integration patterns, prompt engineering, and modern full-stack practices.
Stay current with AI/ML tooling, LLM advances, and MCP ecosystem developments — bring what you learn back to the team.