AI-Directed Software Engineer

Job Summary: Job Summary: The AI-Directed Software Engineer designs and delivers software across the full Envisionware stack (backend, frontend, native clients, and the AWS infrastructure that runs them) by directing AI systems to do the bulk of the implementation work. You'll move ideas from concept to demo to production at high speed, decomposing problems into spec-driven AI-executable tasks, steering and refining AI output, and owning the quality, performance, and customer impact of the result. This is a full-stack role with a DevOps component, built on AWS. You won't be specializing in one layer. You'll own features end-to-end: from the database schema, through the Java services, through the Angular or React UI, and out through the Kubernetes deployment on AWS that ships them. You'll also build AI capabilities into the platform itself, using Amazon Bedrock to deliver agent-driven workflows on AWS. Operating in an AI-first environment, you'll push the organization from AI-assisted toward AI-delegated software delivery. Job Description: Job Description: Our Stack AWS experience is required. The rest you can ramp on, but you'll work across all of it: • Backend: Java, Maven, Jersey (JAX-RS), Jackson, Log4j, Tomcat • Frontend (Angular): Angular + TypeScript, PrimeNG, PrimeFlex, Transloco, DayPilot • Frontend (React): React, Vite, TypeScript, TailwindCSS • Data: PostgreSQL (runtime + analytics instances), direct SQL • Messaging: Apache ActiveMQ • Native client: C# / .NET Framework (Windows ZeroClient), WiX / MSI installers, Android/iOS • Cloud (AWS, required): CloudFormation, IAM, VPC networking, S3, EFS, CloudWatch, multi-tenant cloud architecture • Containers & DevOps: Docker, Kubernetes, Helm/Kustomize • AI Platform: Amazon Bedrock (foundation models, Agents / AgentCore, Knowledge Bases, Guardrails), AWS Lambda for agent tools and action groups, Claude Code • Analytics & tooling: Python ETL pipelines, Swagger / OpenAPI Multi-module monorepo: WebServices (backend) · WebApps (Angular) · ReactWebApps (Vite/React) · NativeClients (C#) · Packages (Docker/K8s). What You'll Own • End-to-end delivery of features from concept → demo → production, across backend, frontend, and deployment • Directing AI tools to generate code for spec-driven SDLC, APIs, UI, SQL, infra config, and workflows • Decomposing product requirements into AI-executable tasks • Validation, testing, and hardening of AI-generated output • Kubernetes/Docker configuration and deployment on AWS of the services you build • Designing and shipping Amazon Bedrock-based AI agents: action groups and tool integrations, Knowledge Base (RAG) retrieval, and Guardrails • Operating what you build on AWS, including IAM least privilege, cost awareness (token and compute spend), and CloudWatch observability • Throughput and cycle time across your assigned workstreams • Continuous improvement of AI-driven development patterns, prompts, and tooling How AI Changes This Role AI is your primary implementation engine. You're not expected to hand-write every line of code across every layer of this stack; you're expected to direct AI to produce it. You'll use AI to generate Java services, Spring Boot, Angular components, React UIs, SQL, Dockerfiles, K8s manifests, and CloudFormation templates alike. AI is also part of what you'll ship. Bedrock-based agents will put AI to work inside our products, which raises the bar: agents need guardrails, evaluation, human-in-the-loop checkpoints where decisions matter, and predictable cost. Every AI-generated output is a starting point, not a finished product. You own correctness, edge cases, security, and production readiness. The breadth of this stack is exactly why AI-directed development matters here: no single engineer can be a deep expert in Java, Angular, React, C#, PostgreSQL, Kubernetes, and AWS, but one engineer directing AI across all of them can. What We're Looking For Required • Strong software engineering fundamentals (APIs, distributed systems, debugging, data flows) • Hands-on AWS experience (3+ years) building and running production workloads; able to reason about IAM, VPC networking, and the cost and failure modes of the services you use • Infrastructure as code on AWS (CloudFormation preferred; CDK or Terraform acceptable) • Full-stack breadth: comfortable moving between backend services, UI, and deployment config in the same day • Working familiarity with containers and Kubernetes (or willingness to ramp fast); can debug a failing pod, read a manifest, and ship a Helm change • Demonstrated experience using AI coding tools (Claude Code, Cursor, Copilot, or similar) to ship real work • Sharp eye for reviewing AI output, especially subtle correctness, security, or deployment issues • Comfort in fast, ambiguous, rapidly changing environments • Bias toward shipping working software over perfect design • Systems thinking: understanding how components interact at scale • Willingness to challenge both human and AI-generated assumptions • Strong written communication: prompting is writing Strongly Preferred • Hands-on Amazon Bedrock experience: invoking models from code, building agents with tool use / action groups, RAG with Knowledge Bases, and applying Guardrails • Evaluating and validating LLM and agent output in production (test harnesses, eval sets, failure handling), not just demos • Familiarity with agent patterns more broadly (MCP, multi-step tool orchestration, human-in-the-loop checkpoints) Nice to have: Experience with Java/Jersey, Angular or React, PostgreSQL, or library, public-sector, or multi-tenant SaaS domains. What Success Looks Like (First 90 Days) • Ship multiple features from concept to demo-ready in ≤5 days each, touching backend, frontend, and deployment where required • Demonstrate effective use of AI to produce production-quality code across the stack • Deliver at least one Bedrock-backed agent workflow from prototype to int