Senior Full Stack AI Agentic Developer

About Capnexus

Capnexus is a comprehensive services provider. Our team consists of outstanding professionals, highly experienced in designing, building, and supporting retail software. We see ourselves as a build-as-a-service provider who follows a repeatable business pattern that can be applied to a variety of platforms and verticals. With a culture built on outcomes and delivery, Capnexus provides customers with a complete suite of services for software development, system analysis, integration, implementation, and support — including the option to engage a single team to perform all the services they require.

Who You Are and What You’ll Do

Capnexus is looking for a highly skilled Senior Full Stack Developer to design, develop, and deliver a cloud-native application for an AI-powered modernization engagement. This is an opportunity to work on cutting-edge generative AI solutions.

Key Deliverables

• Self-service subcontractor portal: a static Next.js/React SPA on S3 + CloudFront, with real-time data input verification and submission status tracking
• Automated subcontractor submission flow with smart forms, document uploads and an embedded assistant
• Risk Profiler dashboard covering subcontractor risk ratings, submission tracking and AI-driven follow-up
• Exception Management dashboard with flagged subcontractors, risk scores and routing/approval workflows
• Python Lambda backend behind API Gateway REST, built to an OpenAPI contract, including Autodesk and CMIC ERP integrations
• Aurora PostgreSQL Serverless v2 data layer behind RDS Proxy, with Flyway migrations, plus DynamoDB for session and job state
• Identity and RBAC: Cognito with Entra ID SAML federation and a group-derived capability model for multi-party access
• AWS CDK v2 (TypeScript) infrastructure repo, with cdk-nag as a synth gate and GitHub Actions OIDC deploys across four environments
• Up-to-date specs, plus technical documentation: architecture diagrams, deployment guides and operational runbooks
• Security implementation covering encryption at rest and in transit, VPC design, IAM policies and audit logging

Responsibilities

• Work spec-first inside the AutoMod workflow. Contribute to requirements, architecture and per-epic implementation specs in the discovery repo using Kiro (specs, steering files, hooks).
• Review agent-generated code and specs closely and give precise, actionable feedback
• Trace defects back to their source in the requirements or specs, fix them there, and make sure each fix carries through every affected epic
• Pass work through the automated gates (route manifest, coverage, cdk-nag, schema parity) and resolve items in the formal findings loop
• Build the Python 3.14 Lambda backend with Powertools, SQLAlchemy 2, Pydantic 2 and PyJWT, behind API Gateway with a Cognito JWT authorizer
• Build the Next.js 16 / React 19 / TypeScript frontend with Tailwind, shadcn/ui (Radix), TanStack Query and Zustand
• Develop against an OpenAPI contract using MSW/Prism mocks. Test with Vitest, and with Playwright + axe for end-to-end and accessibility checks.
• Design and maintain the Aurora PostgreSQL schema and Flyway migrations, the DynamoDB state stores, and S3 + EventBridge ingestion
• Implement the Cognito/Entra ID federation and capability-based RBAC for multi-party portal access
• Integrate the CMIC and Autodesk ERP APIs, SES-based subcontractor outreach, and Amazon Bedrock-driven document validation and exception workflows
• Maintain the CDK infrastructure and GitHub Actions pipelines across all four environments
• Support integration testing and UAT, write training materials, and contribute to knowledge transfer and post-deployment operations guides
• Work with the AI/ML, DevOps and Data Engineering team members

Qualifications

• Hands-on experience with spec-driven, agentic software development: working spec-first rather than code-first. Kiro (CLI/IDE, steering files, specs, hooks) is strongly preferred; experience with an equivalent agentic tool is also considered.
• Proven ability to review AI-generated output critically and drive corrections back into the requirements and specs rather than patching code
• 5+ years of full stack development, including 2+ years in AWS serverless environments
• Strong AWS serverless skills: Lambda, API Gateway, Aurora/RDS Proxy, DynamoDB, S3, CloudFront
• Infrastructure as Code with AWS CDK in TypeScript; CI/CD with GitHub Actions or an equivalent
• Strong Python backend …