Staff Software Engineer, Applied AI

Ruxlo•Wall Township, NJ

2d•$200,000 - $240,000

About The Position

Ruxlo is hiring a Staff Software Engineer, Applied AI to set the technical strategy for incorporating AI into its product and serve as the technical anchor for applied AI across teams. The role spans LLM-powered features, retrieval, structured data extraction, and long-running autonomous agents that can use tools, pause for human approval, and safely resume after failures. Some agents will act on users’ behalf, so the role requires careful consent, authorization boundaries, and reviewable records of actions. The engineer will own Ruxlo’s AI architecture end to end, ship AI product capabilities and background agents, establish evaluation and observability practices, and define standards for safe AI use. They will also make build-versus-buy decisions, develop shared AI tooling and Model Context Protocol servers, and mentor engineers. Ruxlo notes that production-agent patterns are still emerging and expects this person to set direction amid ambiguity. This is a dedicated, in-house role. Ruxlo is not considering fractional, part-time, or consulting arrangements. Candidates must be authorized to work in the United States; visa sponsorship is not offered. Ruxlo LLC is an Equal Opportunity Employer. Ruxlo may use AI-assisted tools to support application screening, review, organization, scheduling, and communication; final hiring decisions are made by humans.

Requirements

At least 8 years of engineering experience, including a technical lead role on systems used by other engineers.
Experience shipping LLM-powered features to real users, such as retrieval-augmented search, extraction, or classification, and ability to explain what worked, what failed, and how results were measured.
Experience taking at least one autonomous, non-conversational agent to production that was triggered by events, schedules, or data and completed multi-step work without a person in a chat window.
Experience building agents that act for users in other systems, including delegated permissions, consent and confirmation flows, and undo.
Deep command of prompt design, structured outputs, tool use, Model Context Protocol, embeddings and vector search, and context-window management.
Track record of evaluation and observability for AI in production, including tracing, cost metering, and quality measurement without relying on a human to grade every result.
Strong backend fundamentals in Python or TypeScript, PostgreSQL, and distributed systems on AWS with Docker and Kubernetes.
Excellent verbal and written communication skills.
Ability to work independently and as part of a team.
Adaptability and comfort working in ambiguity.
Strong ownership and follow-through.
Effective use of AI tools in daily work.
Sound decision-making and prioritization.
Openness to feedback and ability to give feedback well.
Ability to collaborate across functions.
Authorization to work in the United States; visa sponsorship is not offered.

Nice To Haves

Hands-on experience with durable execution or workflow systems; Temporal is strongly preferred, and Restate, Inngest, AWS Step Functions, or similar also count.
Experience with agent frameworks and runtimes such as the OpenAI Agents SDK, Claude Agent SDK, LangGraph, or OpenClaw.
Classic machine learning experience for entity resolution or record matching over large structured or geospatial datasets.
Experience fine-tuning or distilling smaller models for high-volume, narrow tasks.
Cost engineering at scale, including model selection per task, prompt caching, batching, and token budgets.
Open-source contributions to agent runtimes, workflow engines, or Model Context Protocol servers.
Experience at a founder-led or bootstrapped startup.

Responsibilities

Own the architecture of Ruxlo’s AI systems, including model access and routing, retrieval, prompt and tool layers, agent runtime, evaluation, and cost.
Ship LLM-powered product features, including natural-language search, summarization, structured extraction, and classification over property, ownership, and contractor records.
Ship autonomous background agents that verify, reconcile, enrich, and keep records fresh.
Build agents that act on users’ behalf, including requesting quotes, following up with contractors, importing work history, and filing documents, with user consent at consequential steps.
Design durable, long-running agent workflows using an execution engine such as Temporal, including retries, timeouts, idempotent tool calls, and safe resumption after failure.
Establish evaluation for every AI feature, including golden datasets, offline evaluation, regression gates in continuous integration, and production sampling.
Set standards for safe AI use, including prompt-injection defense, least-privilege tool access, delegated authority limited to what users could do themselves, and reviewable audit trails.
Make build-versus-buy decisions across model providers, frameworks, and vector stores while treating cost and latency as first-class constraints.
Build shared AI tooling, skills, and Model Context Protocol servers.
Mentor engineers on building with AI.