Backend Developer Needed for HAI Automation Hub Completion

I am looking for a strong backend developer who can work quickly and efficiently using AI coding tools such as Codex, Cursor, Claude Code, or a similar agentic coding environment. The project is a local-first automation hub / personal AI operating system. It uses a Go backend, Angular frontend, Docker Compose, Postgres, Redis, Kafka, nginx, workflow automation, memory, verification, connected sources, and guarded agent runtime adapters. A lot has already been built, but there are still important missing, incomplete, or not fully verified parts. I want the remaining backend and infrastructure work pushed much closer to a genuinely working local product. ## Goal The goal is to complete, harden, and verify as much of the missing backend and infrastructure work as possible. This is not a “start from scratch” project. The developer should inspect the existing repository, understand the architecture, identify the highest-impact gaps, implement improvements, and produce a clean GitHub pull request. ## Main priorities The work should focus on: 1. Backend completion and hardening. 2. Full Docker Compose readiness. 3. Health/readiness visibility. 4. Connector and adapter readiness. 5. UI/UX visibility for what the system is doing. 6. Realistic verification evidence. 7. A clean pull request that can be reviewed and merged. 8. Remote verification on the actual laptop using TeamViewer or AnyDesk. 9. Clear Trello-based progress updates. ## Known areas that need attention The developer should inspect the repository first, but known gaps include: * The full Docker Compose stack has not yet been proven end-to-end. * Postgres + Redis + Kafka + nginx + backend + frontend should boot together reliably. * `/healthz` and `/readyz` should clearly show what works, what fails, and what is missing. * Connector status should be visible and honest. * Gmail, Drive, Calendar, GitHub, Trello/project-board style connectors are mostly not implemented yet. * The normalized JSON-feed bridge should be made practically usable where full OAuth connectors are not realistic yet. * OAuth/webhook callback readiness needs improvement. * LLM provider generation is not fully exercised by default. * Vector embedding infrastructure is not production-grade yet. * Quota accounting across restarts is not production-grade yet. * Secret management, migrations, RBAC hardening, and deployment configuration need improvement. * Several tested utility packages exist but are not wired everywhere. * The UI should make system state visible: readiness, connector status, provider status, blocked actions, missing configuration, running jobs, failures, and recommended next actions. ## Existing infrastructure Where applicable, I want to use existing infrastructure instead of rebuilding unnecessary alternatives. For example: * An ngrok subscription is already available. * ngrok can be used where a temporary public URL is needed for webhook callbacks, OAuth callback testing, external connector testing, or remote verification. * The project should still remain local-first and secure. * Public exposure should go through the intended gateway/reverse-proxy path where possible, not by exposing unsafe backend services directly. * Any ngrok usage must be documented clearly: what it connects to, what it was used for, and whether it is required permanently or only for testing. ## Trello communication requirement The developer will be granted access to a dedicated Trello card for this job. All project updates must go through that Trello card. This is required so progress, blockers, decisions, links, verification notes, and final delivery remain traceable in one place. The developer must use the Trello card for: * daily progress updates; * questions or blockers; * links to branches, commits, and pull requests; * notes about what was changed; * notes about what was tested; * notes about what failed; * remote verification findings; * final delivery notes. Important: if something is discussed elsewhere, the outcome must still be summarized on the Trello card. The Trello card is the source of truth for the job. ## Remote verification requirement This job includes logging in to the laptop remotely via TeamViewer or AnyDesk in order to verify that the tool actually works in the real local environment. This is important because the project is local-first and depends on local Docker, local configuration, local connectors, and local runtime behavior. The developer must be willing to: * join a remote session using TeamViewer or AnyDesk; * help run the project on the laptop; * verify Docker Compose startup; * verify dashboard access; * verify health/readiness status; * verify connector behavior where possible; * verify ngrok/webhook/OAuth callback behavior where applicable; * troubleshoot local environment issues; * document what was tested on the machine; * clearly separate “verified on the local laptop” from “only tested in the developer’s own environment.” The developer should not make uncontrolled changes to the machine. Any sensitive or destructive action must be discussed first. ## Expected deliverables By the end of the job, I expect: 1. A GitHub branch and pull request. 2. Clear commits with meaningful commit messages. 3. Backend improvements for the highest-impact missing areas. 4. Any necessary Angular UI/UX additions so I can see what is happening. 5. Updated environment example and setup instructions where needed. 6. A verified local run path using Docker Compose. 7. Remote verification on the laptop through TeamViewer or AnyDesk. 8. Practical connector verification where possible. 9. ngrok setup or documentation where it is useful for callbacks, webhooks, OAuth testing, or external access. 10. Tests or smoke checks for the implemented work. 11. Verification notes showing exactly what commands were run and what passed or failed. 12. Trello card updates documenting progress, blockers, decisions, verification, and final delivery. 13. No fake success, no mock production claims, and no pretending something works if it was not actually tested. 14. A final report containing: * what was completed; * what remains; * what was verified locally by the developer; * what was verified on the laptop; * what commands were run; * what passed; * what failed; * what still needs to happen next. ## Important expectations This is an AI-assisted development task. I specifically want someone who can use Codex, Cursor, Claude Code, or another strong AI coding workflow to move faster than ordinary manual development. You do not need to manually code everything line by line. I am looking for someone who can: * guide AI coding agents effectively; * review generated code critically; * fix mistakes; * run tests; * verify behavior; * avoid fake completion; * produce a clean pull request; * help verify the project on the actual laptop through remote access; * keep all updates traceable through the Trello card. You must be comfortable with: * Go backend development; * Docker Compose; * Postgres; * Redis; * Kafka or event-driven systems; * nginx reverse proxy setup; * ngrok or similar tunneling tools; * webhook and OAuth callback testing; * API design; * authentication and authorization; * local-first software; * Angular basics for UI/UX visibility; * remote troubleshooting through TeamViewer or AnyDesk; * Trello-based project communication; * debugging existing repositories; * writing clear verification notes. ## Budget Fixed price: **$35** ## Deadline The work should be completed within **3 days after hiring**.