phoenix
phoenixPhoenix - Your AI Assistant
Inactive session will be ended in 24 hours
End Session

Full-Stack Developer for AI SEO & GEO Workflow Platform

  • Posted:
  • Proposals: 63
  • Remote
  • #4508833
  • OPPORTUNITY
  • Open for Proposals
Aadil S.Dhawal T.Salihu G.Yasir Ali M.Mary O. + 58 others have already sent a proposal.
  • 12

Description

Experience Level: Expert
We are looking for an experienced full-stack developer to help complete Tessy, an internal AI-powered platform for managing SEO and GEO client delivery.

The frontend has already been built in Lovable and is being moved into Replit. We now need a developer to review the existing code, complete the application architecture and build the backend, database, integrations and production workflows.

Core Phase 1 Features

* Secure login and role-based access for admins, account managers, delivery staff and clients
* Client onboarding and account setup
* Separate but connected SEO and GEO workflows
* AI-generated strategy drafts based on client data and templates
* Automated “Critic” review before human approval
* Inline editing and approval by account managers
* Automatic generation of a rolling three-month task plan
* Task assignment, statuses and review workflows
* Client approval and delivery through Tessy
* PDF or portal-based deliverables
* Logging of publication dates and live URLs
* Recording account manager edits as before-and-after data
* Capturing performance results two and four weeks after publication

Tessy should manage the delivery workflow from onboarding through to release and performance tracking. The detailed PRD and Lovable prototype will be provided.

Developer Responsibilities

* Review the Lovable and Replit codebase
* Recommend a suitable production architecture
* Build the backend and database
* Connect the existing frontend to live data
* Integrate AI APIs and structured outputs
* Implement authentication and user permissions
* Build the workflow, task and approval logic
* Add file storage and PDF generation
* Test, deploy and document the application

We are open to recommendations on the final stack. Relevant experience may include React, TypeScript, Supabase, PostgreSQL, Node.js, OpenAI or Anthropic APIs, Lovable and Replit.

When Applying

Please include:

* Examples of similar applications you have built
* Your experience taking AI-generated prototypes into production
* Your experience with Lovable and Replit
* Your recommended technical approach
* Your availability
* An estimated timeline and budget
* Any major risks or questions you would raise before starting

Please begin your proposal with “Tessy” so we know you have read the brief. We prefer milestone-based delivery with regular demonstrations and access to the working code.

NB: We have a PRD if requested and can provide access to loveable FE.

New Proposal

Login to your account and send a proposal now to get this project.

Log in

Clarification Board Ask a Question

  • Gabriel A.

    Regarding the existing codebase
    How complete is the current Lovable → Replit migration? Is the frontend primarily a polished UI prototype, or has any backend integration and state management already been implemented?

    Regarding the AI workflow
    For the AI-generated strategy drafts and the "Critic" review, do you already have prompts, templates, and evaluation criteria defined in the PRD, or are you looking for the developer to help design that architecture as well?

    Regarding scalability
    Is Phase 1 expected to support multiple client organisations (multi-tenant SaaS), or will Tessy initially operate as a single-company internal platform with the option to expand later?

  • CYTNHIA W.

    The workflow captures AI drafts, critic reviews, account manager edits, and client approvals. Do you expect every transition to be fully event-sourced with an immutable audit trail, or is versioning at the document level sufficient for your compliance requirements?
    How do you want to prevent AI outputs from becoming inconsistent over time? Should prompts, templates, critic rules, and model versions all be version-controlled so historical deliverables remain reproducible months later?
    The Lovable frontend was built as a prototype. Have you already assessed whether its component architecture is production-ready, or are you expecting the successful developer to identify and replace parts that won't scale before backend integration begins?
    When a client requests changes after approval, should the system branch into a new workflow while preserving the original deliverable, or should approvals remain editable within the same workflow instance?
    How do you envision scaling the AI pipeline as your client base grows? Should strategy generation, critic evaluation, PDF creation, and performance tracking be processed synchronously, or would you prefer an asynchronous job queue to avoid blocking user workflows?
    Will SEO and GEO remain independent workflows with shared data, or do you ultimately see them evolving into configurable workflow templates that different service lines can reuse without code changes?
    How important is explainability of AI decisions? Should account managers be able to see why the critic rejected a strategy, or only receive the revised output?
    What is the biggest pain point with the current Lovable/Replit implementation: is it missing functionality, architectural limitations, or concerns about taking generated code into production?

  • Yasir Ali M.

    Hello William, regarding the Lovable prototype, are there specific API endpoints or data structures already defined that I should align with, or should I propose a fresh schema for the PostgreSQL database?

  • Khalid H.

    would like to clarify a few points:

    1. Is Supabase your preferred backend, or are you open to alternatives such as PostgreSQL with NestJS or Node.js?
    2. Will AI prompts and templates already be available, or should they be designed as part of the project?
    3. Do you already have deployment infrastructure in place, or should I also handle production deployment?
    4. Will client notifications (email/in-app) be required during approval workflows?
    5. Is Phase 1 expected to be fully production-ready, or is an MVP acceptable initially?