AI Agent Infrastructure Engineer — Agentic Harness, Daytona Sandboxes, Parallel Agents-ASAP

Title: AI Agent Infrastructure Engineer — Agentic Harness, Daytona Sandboxes, Parallel Agents (URGENT — Days, Not Weeks) Overview Healthcare SaaS company, team of ~10 engineers, latest .NET / Angular / MS SQL stack. We've already deployed a multi-model agent stack — Claude Code, LiteLLM gateway, Git worktrees, MCP integrations — and we need someone who has shipped this kind of infrastructure end-to-end before. Not someone learning on the job. This needs to move fast. If your timeline is measured in weeks, don't apply. We want someone who looks at this scope and gets it done in days. We're open to fixed price per milestone if that's how you work best. This starts as a focused engagement but we expect ongoing work — follow-ups, fixes, enhancements as the AI tooling landscape keeps shifting. If you're the right person, this turns into a recurring relationship. What You Need to Have Actually Done (not just read about) Built full agentic coding harnesses — orchestrator → subagent → CI gate → merge loops, end to end Deployed Daytona (or equivalent: E2B, Modal) as an isolated sandbox execution layer for AI-generated code Run parallel Claude Code agent sessions via Git worktrees — multiple simultaneous subagents on scoped tasks Built or configured a self-hosted LiteLLM gateway routing to multiple models (Claude, GPT, Gemini, DeepSeek, Kimi or similar) Wired up MCP servers (file system, SQL, Atlassian, Slack, or similar) as tool layers for agents Structured agent harnesses using CLAUDE.md, COMMON_MISTAKES.md, subagent role definitions, hook scripts, and settings.json — or a directly comparable framework Implemented Plan Mode or equivalent human-in-the-loop approval gates before agent execution Built something like the 7-agent feature factory pattern (Researcher → Spec Writer → Builder → Validator → Verifier) or Dynamic Workflows / multi-agent orchestration equivalent Used Temporal, n8n, or similar for durable long-running agent workflow execution Treated CI green as the mechanical enforcement gate for agent output quality Our Current Stack (what you're plugging into) Latest .NET / C# backend, latest Angular frontend, MS SQL, EF Core Docker Compose on Windows Server / IIS Bitbucket + Bitbucket Pipelines for CI/CD Claude Code with shared .claude/ directory, CLAUDE.md, COMMON_MISTAKES.md already in place LiteLLM gateway deployed in Docker Compose MCP: Atlassian (Jira/Confluence), custom MS SQL MCP server, Zulip Temporal for durable workflows n8n for automation pipelines Playwright + Autonoma for QA automation (in progress) Scope of Work Audit current agent harness and identify gaps against a full-loop, production-grade standard Configure and deploy Daytona sandboxes as the isolated execution runtime for agent code runs Harden the parallel agent workflow — Git worktrees, subagent role configs, hook scripts, settings lockdown Wire Daytona into the existing CI pipeline so agent-executed code runs in clean snapshots, not against live infra Deliver a runbook / documented standard that our backend lead can own going forward Engagement Structure Fixed price per milestone or hourly — your call, we're flexible Start immediately, deliver fast — this is not a slow burn project Ongoing as needed after initial delivery: fixes, enhancements, and keeping the stack current as models and tooling evolve. AI is moving fast and we need someone plugged in who can adapt the harness as things change How to Apply Skip the generic pitch. Show me something real: A GitHub repo, diagram, or Loom walkthrough of an agentic harness you've actually shipped Specific tools you've configured — Daytona, E2B, LiteLLM, Claude Code, etc. — not names you recognize, things you've touched One sentence on the hardest thing you solved to get full agent loops running reliably Your realistic timeline to close this scope Proposals without evidence of prior work won't be reviewed.