Experienced Software Engineer — AI Operations & Platform Development

Software Engineer - AI Operations & Platform Development Company SilverStay, Inc. (Odenton, MD) Reports To Chief Technology Officer Location Remote (United States — all states eligible) Employment Type Full-Time Compensation $105,000 – $125,000 base, commensurate with experience About SilverStay SilverStay is building an AI-powered hospital discharge management platform that replaces the spreadsheets and printed census tools used by care navigators today. The platform ingests hospital data feeds, applies AI to surface discharge barriers, assesses patient readiness, and recommends next steps across persona-specific views for care navigation teams. We are partnered with Adventist Health, the University of Maryland Medical System, and a growing number of hospital networks. This is a HIPAA-regulated environment built on a serverless AWS stack with a Next.js front end. Role Summary We are hiring a Mid-Level Software Engineer who builds things, manages things, and supports things — all under one roof. This is not a specialized role. The successful candidate will own platform development, build and maintain AI automation workflows that run the internal team’s operations, and serve as the primary Claude AI and tooling resource for the entire company. This role is ideal for an engineer who is equally comfortable writing production-grade AWS infrastructure and prompting an agentic pipeline — someone who has built something real on their own and wants a position where that instinct to build is the job, not a side project. The successful candidate will take over the SilverStay platform codebase as the current development team completes its engagement. You will inherit and expand a set of internal AI workflows already in production. And you will be the person every SilverStay team member turns to when they need to get more out of Claude. Key Responsibilities Platform Development (~40%) Own ongoing maintenance, bug fixing, and feature development of the SilverStay platform following the current development team’s handoff. Work full-stack: Next.js 14+ / React / TypeScript front end and AWS serverless back end (API Gateway, Lambda, RDS/PostgreSQL, S3, Cognito). Deploy and maintain infrastructure using AWS CDK; follow established stack standards, tagging requirements, and deployment profiles. Participate in code review, testing, and CI/CD workflows; keep the codebase healthy, documented, and auditable. Contribute to the platform’s pluggable analysis engine — adding and tuning AI analysis jobs as new hospital requirements emerge. AI Operations & Agentic Workflow Engineering (~40%) SilverStay’s internal team handles high-volume, time-sensitive work: qualifying calls, facility placement, referral outreach, family communications, contract patient documentation. This part of the role is about building and maintaining the AI-powered workflows that take those tasks off the team’s plate. Own existing AI automation workflows — already in production — and maintain, improve, and extend them. Active workflows include structured call summaries from qualifying call recordings; facility match identification and auto-drafted referral emails; agentic tracking of referral responses; and hospital-specific patient summaries for contract cases. Build lightweight agentic workflows that monitor inputs, take action, and loop back for human review — without requiring step-by-step instruction. Extend agentic pipelines using Claude and the Anthropic API; write, test, and refine prompts that produce consistent, auditable outputs in a clinical context. Track AI usage and cost across internal workflows; optimize for prompt efficiency; identify caching and batching opportunities. Document every workflow so it is repeatable, auditable, and not dependent on a single person. Claude Desktop Support & Internal AI Enablement (~20%) Every SilverStay team member uses Claude in their daily work. You are the person who makes that work well. Serve as the company’s primary Claude Desktop resource: configure, troubleshoot, and extend Claude’s usage across the team. Provide hands-on support when team members run into limitations, get poor outputs, or need guidance on how to structure their prompts and workflows. Build and maintain shared prompt libraries, templates, and workflow guides that scale your support across the team without requiring 1:1 help every time. Use Claude Code to accelerate your own development work and serve as an internal reference for how to do the same. Stay current on Anthropic product releases, model updates, and new capabilities — and proactively surface what’s relevant to SilverStay operations. Required Qualifications The following are hard requirements. Applications that cannot demonstrate these will not advance. 3–5 years of professional software engineering experience (internships and co-ops may count toward this if accompanied by strong independent project work). AWS — not GCP, not Azure. You must have hands-on, production experience with core AWS services: Lambda, API Gateway, RDS/PostgreSQL, S3, IAM, and Cognito. AWS CDK or other infrastructure-as-code experience is strongly preferred. Full-stack engineering. You work comfortably on both sides: front end in Next.js / React / TypeScript and back end in AWS serverless. You do not specialize in one and tolerate the other. Agentic AI experience. You have built at least one agentic or LLM-powered workflow in a production or near-production context — not a demo, not a tutorial. You understand tool-calling, prompt chaining, and the failure modes that come with autonomous AI systems. Claude and Anthropic API experience. You have used Claude Desktop for real work and written code against the Anthropic API. Experience with Claude Code is a significant plus. Independent builder. You have created something from scratch on your own time: an application, a tool, a data project, an automation — something you built because you wanted to, not because someone assigned it. You do not need to have published it, but you need to be able to describe it, show it, or walk us through it. Strong written communication. You document your work clearly for both technical and non-technical stakeholders — this matters in a clinical environment where auditability is non-negotiable. Authorized to work in the United States. Preferred Qualifications Experience with AWS Bedrock or other managed AI/ML services on AWS. Familiarity with agentic frameworks such as LangChain, LangGraph, or similar orchestration tooling. Experience inheriting and stabilizing a codebase from an outside development team. Exposure to SOC 2 or other compliance-driven environments beyond HIPAA. Experience with single-table DynamoDB design and serverless architecture patterns. Active GitHub profile, portfolio, or public project work that demonstrates what you build on your own. HIPAA-regulated environment. You understand least-privilege access, no PHI in logs, and audited actions — and you treat these as engineering standards, not compliance checkboxes. Interest in healthcare technology and improving care outcomes — this problem is worth solving, and it helps if you believe that. Compensation & Salary Range This role is compensated at mid-level market rates for the full skill stack required, including the AWS, full-stack, and applied AI/agentic premium that the 2026 market reflects. Minimum Midpoint Maximum $105,000 $115,000 $125,000 Base salary only. Range reflects remote US market rates for mid-level full-stack engineers with applied AI and AWS specialization as of mid-2026. Benefits, equity participation, and professional development budget discussed during offer stage. What Success Looks Like — First 6 Months Clean knowledge transfer from the outgoing development team, with no critical coverage gaps. Independent ownership of the deployment pipeline and day-to-day platform maintenance within 60 days. AI automation workflows running reliably, with issues caught early and costs tracked against budget. At least one new hospital configuration or platform feature shipped end-to-end. A Claude Desktop support system in place that reduces ad hoc questions and scales your support across the team. At least one new agentic workflow designed, built, and handed off to the team for daily use. Show more Show less