Founding AI Engineer - Business Automation

Blueline Search Group•Washington, DC

21h•$140,000 - $175,000•Hybrid

About The Position

The company we are recruiting for is a public accounting firm focused on audit, tax, and advisory services for the insurance and nonprofit sectors, as well as employee benefit plans. For more than three decades, they’ve combined deep domain expertise with a people-first culture built on agility, respect, and trust. They’re investing heavily in technology to transform how assurance and tax services are delivered—so their teams can spend more time on high-value judgment and our clients can benefit from faster, more insightful outcomes. You’ll join our Business Automation team, which is chartered with building the next generation of tools that power how we work—data, automation, and AI/agentic systems. This is our first AI engineering hire and a truly greenfield role: no legacy codebase, no pre-baked platform. You will: Act as a founding AI engineer and primary builder of our AI stack on AWS Design and deliver our vision for an agentic workforce—AI systems that can reason, orchestrate tools, and collaborate with our people Lay the groundwork to grow into a management/tech-lead role, mentoring future AI and application engineers We are deeply invested in AWS and use multiple model providers (primarily AWS Bedrock, but also OpenAI, Gemini, Grok, and others). We expect you to select and combine the right tools for each use case—balancing risk, cost, performance, and long-term maintainability. You’ll also be the technical owner of key AI vendor relationships, working with partners who have helped us deliver proofs of concept and turning promising experiments into robust, in-house capabilities.

Requirements

5+ years in software engineering, ML engineering, or applied AI roles, with at least 2+ years building LLM- or NLP-centric applications.
Strong experience building on AWS, including secure production workloads using services such as Lambda, API Gateway, Step Functions, ECS/EKS, S3, CloudWatch, and IAM.
Hands-on experience shipping AI/LLM applications to production, such as: Retrieval-augmented generation (RAG) systems over private data
Tool-using agents / function-calling workflows
Document understanding / extraction pipelines
Practical experience with multiple model providers (e.g., AWS Bedrock, OpenAI, Gemini, Grok) and a point of view on how to choose the right model for the job.
Strong programming skills in Python (preferred) and/or TypeScript/Node.js for backend services, including API design, testing, and integration.
Experience with embeddings and vector search (e.g., OpenSearch, pgvector, or dedicated vector databases) and building retrieval pipelines with evaluation and monitoring.
Familiarity with security, privacy, and compliance considerations for sensitive data (financial, insurance, healthcare, or similar regulated domains).
Practical CI/CD experience (Git-based workflows, automated testing, deployment pipelines) and comfort with infrastructure-as-code (Terraform, AWS CDK, or similar).
Comfortable operating as a first-of-its-kind hire: setting standards, making build-vs-buy decisions, and delivering under ambiguity with limited pre-existing infrastructure.
Excellent communication skills with non-technical stakeholders; able to facilitate discovery sessions and explain trade-offs in plain language.
Bachelor’s degree in Computer Science, Engineering, or a related field (or equivalent practical experience).

Nice To Haves

Background or experience with audit, tax, insurance, or nonprofit organizations.
Experience designing or operating agent frameworks and workflow orchestrators (e.g., multi-step agent graphs, tool-calling frameworks, evaluator agents).
Experience fine-tuning or adapting models (instruction tuning, RAG-first design, lightweight adapters) in a production context.
Familiarity with data engineering patterns and working with messy real-world data (Excel, PDFs, unstructured documents).
Experience building simple front-end interfaces (e.g., React) to deliver AI capabilities directly to end users.
Prior mentorship or team-lead experience and an interest in building a small team over time.

Responsibilities

Design our AI / agentic foundation on AWS
Define the reference architecture for LLM and agentic workloads (orchestration, retrieval, tools, evaluation, guardrails) on AWS.
Leverage services like Bedrock alongside external providers (OpenAI, Gemini, Grok, etc.) in a modular, provider-agnostic way.
Build AI products that transform core audit & tax workflows
Deliver copilots and agents that help our people draft workpapers, analyze documents, summarize findings, generate testing selections, and prepare client deliverables.
Work end-to-end: discovery with domain experts, prototyping, productionization, monitoring, and iteration based on feedback.
Implement robust RAG & tool-using agents
Design retrieval pipelines over internal content, structured data, and workpapers with appropriate metadata and access controls.
Implement agents that can call tools (internal APIs, calculators, automations, workflows) and operate within well-defined boundaries.
Codify safety, privacy, and governance for AI
Partner with risk, security, and data teams to ensure responsible use of client data (PII handling, redaction strategies, no-train boundaries, access controls).
Establish patterns for prompt hardening, guardrails, and evaluation that reflect our professional standards obligations.
Own AI vendor and partner relationships
Serve as the hands-on technical counterpart for AWS, OpenAI, Google, xAI, and other partners.
Evaluate new capabilities, run proof-of-concepts, and decide when to build vs. buy vs. extend vendor solutions.
Collaborate across data, automation, and domain experts
Work closely with our Data Engineer and automation leads to connect AI systems to clean, governed data and existing workflows.
Translate ambiguous business problems into concrete AI/engineering deliverables and roadmaps.
Establish engineering standards & mentor future hires
Set best practices for AI application development (testing, observability, CI/CD, experiment tracking, documentation).
Help hire, onboard, and mentor junior and mid-level engineers as the AI function grows.