AI Engineer

Raydar; New York (NY); Hybrid; USD 350,000 - 500,000; Full time

Benefits offered by this job

Hybrid in NYC

Competitive equity

Visa transfers considered

Direct ownership of a core product

Job summary

Raydar is seeking an AI Engineer in a hybrid NYC role to own the conversational intelligence layer end-to-end. You will shape prompts, tool calls, and multi-step reasoning to deliver reliable product behavior across agent loops and memory systems.

You’ll productionize stateful user preferences, build personalized ranking and retrieval experiences, and improve system reliability with robust error handling and evaluation frameworks.

Qualifications

• At least four years of experience as a full-stack or backend-leaning software engineer.
• Experience shipping LLM-powered ranking, retrieval, or recommendation systems into production.
• Hands-on ownership of systems where models interpret user preferences and surface results.
• Real evaluation experience, including defining, running, and iterating on quality metrics.
• Comfort across prompting, embeddings, retrieval, memory, preferences, and product-facing agent workflows.
• Strong ability to explain specific production failures you diagnosed and corrected.
• An AI-native development workflow and a track record of learning and shipping quickly.
• Willingness to work regularly from New York City.

Responsibilities

• Own the full agent architecture, including prompts, tool calling, multi-step reasoning, and streaming
• Productionize conversational state and memory systems that interpret user preferences
• Build ranking, retrieval, and recommendation experiences that surface useful, personalized results
• Improve multi-turn reliability through refinement loops, error recovery, and graceful degradation
• Design the rendering layer between agent output and the consumer product experience
• Manage model-provider strategy, caching, structured outputs, and fallback behavior
• Define and operate evaluation systems for ranking, retrieval, and agent behavior
• Debug production failure modes and continuously improve quality using evidence from real users

Skills

Full-stack developer

Backend development

LLM-powered systems

Ranking retrieval recommendations

Evaluation metrics

Prompting embeddings memory

Production debugging

AI-native workflow

Job description

Our client is a 15-person consumer AI company building an agentic travel platform that understands individual preferences and can plan and book real-time travel experiences. Its product connects conversational intelligence with live inventory from major travel and hospitality brands, allowing users to discover and book verified options in one conversation.

We are hiring an AI Engineer to own the conversational intelligence layer end to end. You will turn large language model capabilities into reliable product behavior across agent loops, prompt architecture, tool calling, state, memory, streaming, and error recovery.

What you'll do

• Own the full agent architecture, including prompts, tool calling, multi-step reasoning, and streaming
• Productionize conversational state and memory systems that interpret user preferences
• Build ranking, retrieval, and recommendation experiences that surface useful, personalized results
• Improve multi-turn reliability through refinement loops, error recovery, and graceful degradation
• Design the rendering layer between agent output and the consumer product experience
• Manage model-provider strategy, caching, structured outputs, and fallback behavior
• Define and operate evaluation systems for ranking, retrieval, and agent behavior
• Debug production failure modes and continuously improve quality using evidence from real users

Requirements

• At least four years of experience as a full-stack or backend-leaning software engineer
• Experience shipping LLM-powered ranking, retrieval, or recommendation systems into production
• Hands‑on ownership of systems where models interpret user preferences and surface results
• Real evaluation experience, including defining, running, and iterating on quality metrics
• Comfort across prompting, embeddings, retrieval, memory, preferences, and product-facing agent workflows
• Strong ability to explain specific production failures you diagnosed and corrected
• An AI-native development workflow and a track record of learning and shipping quickly
• Willingness to work regularly from New York City

Bonus points

• Experience in travel, search relevance, personalization, memory, or consumer recommendations
• Experience at a high-growth startup or top technology company
• Proficiency in TypeScript, with Python also welcome
• Significant side projects involving AI agents or production‑grade LLM systems
• A strong academic or equivalent technical signal

Benefits

• Total compensation of $350K-$500K
• Competitive equity
• Direct ownership of a core consumer AI product surface
• Visa transfers, including OPT or H-1B transfers, may be considered case by case
• Hybrid in New York City
• Full-time position
• Close collaboration with a small, fast-moving team