The Role:

This position supports a profitable and rapidly growing enterprise software and circular economy company specializing in IT asset lifecycle management and hardware repurposing to reduce electronic waste across major global corporations.

We are looking for a high-level hybrid profile: part Senior Rails Engineer, part Applied AI Engineer, and part internal AI Advisor. The organization is adopting AI capabilities in a deliberate and measured manner. The candidate will lead the development of production-ready features within a mature Ruby on Rails monolith while establishing the necessary evaluation harnesses, safety guardrails, and operational controls required for systems that impact real-world physical inventory and logistics.

Responsibilities:

• Rails Platform Engineering

  • Build and ship features across a mature Ruby on Rails backend (hosted on Heroku) and two React applications (hosted on Vercel).

  • Work within a mature, layered monolith supporting real-world physical operations (state machines, warehouse workflows, device records, and third-party integrations).

  • Balance AI initiatives with core platform reliability, maintainability, performance, and architecture standards.

• AI Infrastructure, Evaluation & Guardrails

  • Build shared AI infrastructure across products and internal developer tooling.

  • Design AI Evaluation Harnesses: Create benchmarking frameworks and evaluation harnesses to measure model performance against historical company data before deployment.

  • Establish Guardrails: Implement confidence thresholds, input/output constraints, tool usage rules, audit logs, and safe failure/human escalation modes when data is incomplete or ambiguous.

• Operational Domain AI Applications

  • Device Data Normalization & Grading: Build AI-assisted workflows to normalize inconsistent device data, resolve serial numbers, and classify hardware using an A–D rating scale.

  • AI Customer Support: Build automated triage, semantic deduplication, and automated Linear ticket creation for product defects with clear human escalation paths.

  • Historical Retrieval & Reasoning: Implement pragmatic retrieval-augmented generation (RAG) and search patterns over years of physical asset, resale, and processing data.

  • Controlled AI Agents: Construct permissioned agents for specific operational actions (record updates, shipment checks) with strict authorization and audit logging.

• Team Enablement & AI DevOps

  • Introduce AI tools into daily engineering, QA, and DevOps workflows (test generation, code reviews, incident investigation).

  • Guide product leaders and engineers on practical AI use cases while clearly articulating technical limitations and trade-offs.

Requirements:

• Experience: 5+ years of professional production experience in Ruby on Rails, including deep expertise with large monolithic architectures.

• Applied AI in Production: Proven experience shipping LLM/ML features to production using applied patterns (RAG, structured outputs, tool-calling agents, classification).

• AI Evaluation Harnesses: Mandatory experience building evaluation/benchmarking frameworks to measure AI accuracy (a primary indicator of success for this role).

• AI Guardrails & Safety: Demonstrated track record of designing confidence thresholds, human-in-the-loop review workflows, and error mitigation strategies.

• Frontend: Hands-on experience with React to test, debug, and complete frontend integration work.

• Data Hygiene: Experience working with messy or incomplete datasets (entity resolution, data normalization).

• Schedule Alignment: Ability to cover at least 75% of EST business hours.

Great-to-Have:

• Experience fine-tuning or adapting models using proprietary datasets.

• Background working with serial numbers, hardware asset tracking, product catalogs, or physical logistics systems.

• Experience introducing AI adoption strategies inside established engineering organizations.

Mindset and Technical Approach:

• Pragmatic Engineer First: Prefers shipping working, measured software over producing theoretical AI strategies.

• Measurement-Driven: Constantly seeks to measure and prove AI feature accuracy rather than relying on impressive demos.

• Pragmatic Technology Choices: Knows when NOT to use AI and can justify deterministic or traditional software solutions when safer and more effective.