Engineering - Agentic AI Engineer (Junior)

Lumen Dynamics is the bridge between senior care and technology, built to strengthen connection where it matters most. Our all-in-one platform brings together sales, marketing, operations, and engagement tools, empowering senior living communities to work smarter, communicate clearly, and deliver care with heart. Rooted in industry expertise and born from the merger of leading solutions, Lumen Dynamics serves as a unifying force across the senior care space. We help communities across the country streamline processes, enhance resident and family engagement, and stay aligned through every stage of care. That’s why everything we build is designed to support stronger collaboration, seamless workflows, and more meaningful experiences for residents, families, and care teams alike. We are looking for a motivated and technically sharp Junior Agentic (AI) Engineer to join Lumen Dynamics ’s engineering team. This entry-level role is designed for engineers with 1–5 years of experience — or strong recent graduates — who have a genuine interest in building production-grade agentic systems for enterprise workflows. You will work directly with customers and cross-functional teams to design, build, and ship AI-powered features that improve outcomes for senior living communities. From architecting multi-agent workflows to owning the retrieval and eval stack, you will gain hands-on experience across the full AI product lifecycle — with a focus on reliability, compliance, and measurable impact. Responsibilities Agentic Systems & Orchestration Build and deploy agentic systems for enterprise workflows — design and implement AI agents (and multi-agent systems) that reason and retrieve data across complex business processes and take action in enterprise systems. Design and ship multi-step agentic systems — planner/executor, tool-using, multi-agent, and human-in-the-loop — for use cases including onboarding, underwriting, case review, and continuous monitoring. Design orchestration, reasoning, and workflows — architect how agents plan, use tools, and coordinate across complex, multi-step processes. Architect agent graphs in LangGraph (or comparable frameworks — CrewAI, AutoGen, Claude Agent SDK) with explicit state, durable execution, retries, and safe fallbacks. Expose agents to production systems via well-typed tools and MCP servers; treat the tool surface area as a product. Full-Stack Implementation & Integrations Own full-stack implementation and integrations — build across LLMs, APIs, backend systems, and lightweight UIs to deliver complete, working solutions. Build and own the retrieval layer powering our agents: chunking strategies, hybrid search (vector + keyword), reranking, and grounded citation. Design and optimize embedding pipelines and vector indexes using pgvector and OpenSearch. Evaluation, Safety & Reliability Develop agentic harnesses to accelerate development — create evaluation frameworks, toolchains, and workflows that enable rapid iteration and improve system reliability. Own the eval stack: curate golden sets, maintain offline regression suites, implement LLM-as-judge, and run online A/B and shadow evals. Ensure reliability, safety, and production readiness — implement guardrails, validation logic, and fallback mechanisms to ensure consistent and trustworthy behavior in production. Technology Stack Languages Python, Node.js, TypeScript Agent / LLM Frameworks LangGraph, LangChain, Claude Agent SDK, MCP, OpenAI SDK Models Anthropic Claude, OpenAI, open-weight where appropriate Retrieval & Data PostgreSQL, pgvector, OpenSearch, Kafka, Redshift, Redis Infrastructure AWS, Kubernetes (EKS), ArgoCD, Terraform Evals & Observability LangSmith / Langfuse / Braintrust, DataDog Qualifications Education & Experience Bachelor's degree in Computer Science, Data Science, AI/ML, or related field, or equivalent practical experience through projects, research, internships, or professional work. 1–5+ years in software engineering (full-stack or backend), or a strong recent graduate with demonstrable project or internship experience at equivalent depth. Familiarity with LLMs or AI-based systems. Internship or research experience in a production AI or data-intensive environment is a strong plus. Required Technical Skills Proficiency in Python; comfortable with NumPy, Pandas, and Scikit-learn. Hands-on experience with at least one LLM framework: LangGraph, LangChain, Claude Agent SDK, or OpenAI SDK. Understanding of RAG architecture: embedding models, vector databases, hybrid search, and reranking. Familiarity with prompt engineering best practices and awareness of LLM failure modes (hallucination, injection, drift). Working knowledge of SQL and relational databases (PostgreSQL, MySQL, or similar). Familiarity with Git version control and Agile/Scrum practices. Preferred Experience with agent systems, orchestration frameworks (LangGraph, CrewAI,AutoGen), or AI tooling. Exposure to MCP (Model Context Protocol) or building typed tool interfaces for LLM agent