AI Engineer

Data; Union City (NJ); Presencial; USD 120 000 - 150 000; Tempo integral

Resumo da oferta

Data in Union City, NJ seeks an AI Engineer to own AI features from day one, guiding architecture and collaborating with senior engineers on AI strategy. You will build LLM-powered capabilities, integrate data sources, and implement robust evaluation and safety guardrails.

You will work with stakeholders to scope problems, compare model providers, and drive improvements in production AI systems while maintaining high engineering standards.

Qualificações

• 2+ years of Python experience and ML fundamentals.
• Hands-on with foundation models and prompt engineering.
• Experience with vector databases and RAG workflows.

Responsabilidades

• Build production LLM-powered features with prompt orchestration and tool calling.
• Create API integrations and RAG workflows linking data systems to vector search.
• Develop evaluation infrastructure for prompt regression tests and AI evaluations.
• Monitor production AI features for accuracy, latency, and token costs.
• Implement runtime safety and reliability guardrails.
• Evaluate models and tools; guide final build-versus-buy decisions with leadership.
• Collaborate with non-technical stakeholders to scope AI problems.
• Lead code reviews to raise engineering standards.

Conhecimentos

Python

LLM concepts

API integration

Formação académica

Bachelors in CS/Engineering/AI

Ferramentas

LangChain/LlamaIndex

Vector DB (pgvector/OpenSearch)

Descrição da oferta de emprego

You will report to the VP of Engineering and work closely with them and other senior engineers on system design, technical direction, and strategic AI opportunities. You will own AI features from day one, with regular design review and hands-on guidance on the harder architectural calls, and take on more architectural ownership as you build depth in production AI systems.

Responsibilities

• Build and ship production LLM-powered and agentic features – prompt orchestration, tool calling, embeddings, structured outputs – favoring reusable components over one-off solutions.
• Build API Integrations, MCP servers, and retrieval-augmented generation (RAG) workflows that connect our existing data systems to vector search, so that AI features surface and explain our published ratings data accurately and cite the source of each figure.
• Build and run evaluation infrastructure – prompt regression tests, LLM-as-a-judge evaluations, synthetic datasets – that catch feature quality regressions before users do.
• Monitor AI features in production. Track accuracy, latency, token cost, and hallucination rates, then debug and tune prompts based on what you find.
• Implement runtime safety and reliability guardrails that keep AI outputs within quality, compliance, and operational bounds.
• Research and recommend orchestration frameworks, vector stores, and model providers, weighing cost and operational burden against speed. Final build-versus-buy calls are made jointly with the VP of Engineering.
• Work directly with non-technical stakeholders to scope problems and give them honest assessments of what AI can and cannot do.
• Evaluate emerging models and tools, use AI-assisted development extensively, and help raise team engineering standards through code review.

Skillset

• 2-3+ years building or deploying machine learning models or AI-driven applications.

Education

• Bachelor’s degree in Computer Science, Engineering, AI/ML, or a related field, or a relevant certificate program. Demonstrated technical proficiency and relevant industry experience will be considered in lieu of a formal degree.

Required

• Python and ML fundamentals. 2+ years of Python, professional or substantial project-based experience, with a solid grasp of core machine learning concepts and API integration.
• LLM and agentic systems. Hands-on experience with foundation model providers (Anthropic, OpenAI, and similar), prompt engineering, and agentic patterns – using orchestration frameworks such as LangChain or LlamaIndex, or integration standards such as the Model Context Protocol.
• Retrieval and data integration. Practical experience querying and tuning a vector database (pgvector, OpenSearch, or similar) for RAG applications.

Nice to Have

• Managed cloud AI services, such as Amazon Bedrock, are preferred.
• Cloud deployment of application code via serverless or containerized patterns, AWS required.
• Scrum or other Agile development experience.
• Cloud, Gen AI, or machine learning certifications, such as AWS Certified Generative AI Developer, Certified Machine Learning, or Solutions Architect.

How You Work

• Resourcefulness and adaptability. You learn quickly against unfamiliar problems, new ideas, and changing technology, and you find a way through when the obvious path is blocked.
• Accountability and trust. Direct, truthful communication. You keep your commitments, admit mistakes and correct them, and give peers respectful, open-minded feedback.
• Comfort with ambiguity. When requirements are incomplete or still moving, you surface the open questions early and bring options rather than assumptions.
• Translation. You can explain what AI can and cannot do to people who do not build it, and turn their problems into a technical scope, in both directions.
• Coachability. You seek feedback and apply it quickly. You treat design and code review as a learning loop rather than a checkpoint to clear, and you are comfortable saying “I don’t know yet” and going to find out.

Spot any inaccurate information? Have a job to share? Let us know.