AI Engineer

AI Engineer (Document AI and LLM Applications) · Full time, Remote | Document Ingestion and RAG | Azure Preferred · Company Description · Alyra Technology builds AI-enabled SaaS and automation products that help SMBs to enterprise teams move faster, operate leaner, and grow reven ...

Descrição do trabalho

AI Engineer (Document AI and LLM Applications)

Full time, Remote | Document Ingestion and RAG | Azure Preferred

Company Description

Alyra Technology builds AI-enabled SaaS and automation products that help SMBs to enterprise teams move faster, operate leaner, and grow revenue without adding headcount. We create new software in categories that have been underserved or stuck in legacy tooling, especially in document-heavy, regulated environments where accuracy and auditability are non-negotiable. We build on modern Azure architecture and applied AI to deliver systems that are explainable, traceable, and production-ready.

We move fast, ship continuously, and measure impact in days and weeks, not quarters. If you want high ownership, hard problems, and the satisfaction of building products that do not exist yet, you'll fit here.

Role Description

We are hiring an experienced AI Engineer with enterprise grade experience with large, complex PDF ingestion and document understanding. You will own the document AI core across our SaaS products and client deployments: ingestion, extraction, retrieval, grounded answers with citations, and evaluation. You will take messy, real-world documents that break most systems and turn them into reliable, measurable pipelines that power production workflows.

This is not a research role and not a prompt-only AI position. You will ship production-grade services, build the evaluation and quality infrastructure that keeps accuracy high over time, and set the standard for traceability, auditability, and cost-aware performance. You will be accountable for measurable quality: extraction accuracy, citation precision, latency, and cost.

Compensation

Compensation will be discussed and finalized based on experience, scope, and performance expectations. Equity or long-term incentive opportunities may be available for exceptional fits and can be discussed later in the process.

What You'll Do

Core Engineering Responsibilities

• Build real life multi agent AI systems and improve large-scale PDF ingestion pipelines across text, tables, images, and scans
• Develop multi-agent architectures for cross-document analysis and comparison, specifically identifying delta-changes in clauses, endorsements, and schedules across large-scale PDF versions.
• Design layout-aware extraction workflows for complex documents: multi-column, tables, forms, exhibits, endorsements
• Create structured outputs from unstructured documents using schemas, validation, and confidence scoring
• Partner with the team on architecture decisions, failure modes, and production hardening

Azure Cloud Engineering

• Work within Azure-first architectures when implementing ingestion and retrieval systems
• Use Azure services as needed (Blob Storage, Functions, Service Bus) to support scalable pipelines
• Integrate relevant Azure AI services when appropriate (Azure OpenAI, Azure AI Search, Form Recognizer, Foundry, Microsoft Agent Framework (Orchestration SDK))

AI System Implementation

• Familiarity with multi-agent systems, using LangChain, CrewAI, or Microsoft Foundry
• Build GraphRAG and retrieval systems that produce grounded answers with citations and page-level traceability
• Implement hybrid retrieval and reranking, chunking strategies, and document hierarchy modeling
• Build evaluation harnesses: golden datasets, regression tests, hallucination and citation quality metrics
• Optimize latency and cost: batching, caching, indexing strategies, and monitoring cost per document and per query

Collaboration and Code Quality

• Ship clean, maintainable systems with strong documentation and tests
• Work closely with engineering and product to translate requirements into reliable systems

Minimum Qualifications

• 4+ years in software engineering with 2+ years building production AI or document AI systems
• Proven experience building PDF ingestion and document processing systems at scale
• Strong understanding of OCR and layout-aware extraction and their failure modes
• Production RAG experience with measurable evaluation, not just prototypes
• Strong data and retrieval fundamentals: chunking, metadata, hybrid search, reranking
• Comfortable with Python and modern backend development, and shipping production services

Strongly Preferred

• Insurance policy analysis or legal document AI work
• Understanding of policy structures: endorsements, exclusions, schedules, definitions, followed policies, year-over-year comparisons
• Regulated environments where accuracy and auditability matter
• Azure experience: Azure OpenAI, Azure AI Search, Form Recognizer, Blob Storage, Functions, Service Bus
• Vector databases: pgvector, Pinecone, Weaviate, or equivalents

Values and Working Style

• A-player execution: you ship production systems, not prototypes
• Obsessed with correctness, traceability, and repeatability
• Strong ownership and accountability
• Clear communication and practical problem solving under ambiguity
• High standards for reliability, performance, and maintainability