AI SaaS platform — We are looking for a development team

About the project

We are developing an AI SaaS platform for content generation (images and videos). Architecturally, it is a wrapper orchestrator: the frontend sends a request → the backend validates and queues it → the GPU cluster (RunPod Serverless) generates → the result is stored in R2 → the user sees the finished content.

We use ready-made templates (SupaStarter or similar) to speed up development. We already have: Product Brief, HTML-Wireframes, Data Entities, Critical User Flows. We need to assemble this into a working product.

Since this is a niche (AI generation, virtual influencers), the following are critically important: a role-based access control system (RLS) with a "double bottom", credit payments (not direct transactions), and content moderation.

We offer payment in the format of FixPrice or Milestones. It is in our interest to make the platform as quickly as possible without losing quality. In other respects, all conditions are discussed individually during the interview.

Who are we looking for?

We are considering two team options:

Option 1 (preferred):

• Backend Developer (Lead) — Middle+/Senior

• Frontend Developer — Strong Middle

• AI Engineer (ComfyUI) — Middle

Option 2:

• Fullstack Tech Lead (Backend + Frontend) — Senior

• AI Engineer (ComfyUI) — Middle

In both cases, we expect the Backend Lead (or Fullstack Lead) to take on the architecture of the entire system.

Roles and stack

Backend Developer / Lead

What needs to be done:

• Design and implement the API (FastAPI or Supabase Edge Functions).

• Set up the database (PostgreSQL via Supabase): schema, constraints, row-level locking for transactions.

• Implement Row Level Security (RLS) in Supabase: each user sees only their content, admins have full access.

• Implement a credit system with full auditing (balance_ledger — immutable log of all operations).

• Set up queues for asynchronous generation (Redis/BullMQ or pg-boss).

• Integrate payments (Stripe + crypto gateways for high-risk).

• Handle edge cases: race conditions (parallel requests), idempotency (duplicate webhooks), stuck jobs (if the GPU crashes).

• Set up webhooks from RunPod and payment systems.

• Integration with Cloudflare R2 (storing results).

• Role and permissions system: user, premium, admin (different limits, access to models, priority in the queue).

Mandatory experience:

• PostgreSQL: understanding of ACID, transactions, SELECT FOR UPDATE.

• Supabase RLS (Row Level Security): experience in setting up access policies at the database level.

• Experience in designing financial systems or billing (ledger, double-entry, constraints).

• Working with queues (Redis, RabbitMQ, pg-boss).

• Experience integrating payment systems (Stripe API, handling webhooks, idempotency).

• Understanding the architecture of asynchronous systems (callback URLs, retry logic, cleanup jobs).

Stack:

• Python (FastAPI) or TypeScript (Supabase Edge Functions + RPC)

• PostgreSQL (Supabase)

• Redis/BullMQ (or pg-boss)

• Docker

• Cloudflare R2 API

What is already available:

• Data Entities (database schema).

• Critical User Flows (how the system should work).

• Product Brief (requirements).

Frontend Developer

What needs to be done:

• Choose an approach: either take a ready-made template (SupaStarter/ShipFast) or write from scratch on Next.js 14.

• If using a template: remove unnecessary parts (old billing, blog), adapt to our credit system.

• If from scratch: set up Supabase Auth, project structure, basic components.

• Implement UI based on HTML-Wireframes.

• Forms for generation (text, parameters, upload via S3 Presigned URLs).

• Real-time generation status (polling via React Query).

• Dashboard (balance, order history, profile).

• Billing integration (purchase credits, payment history).

• Handle states (loading, errors, edge cases).

• Access levels on the frontend: different UI for free/premium/admin.

Mandatory experience:

• Next.js 14 (App Router, Server Components).

• React Query (for polling and state management).

• Working with forms (React Hook Form + Zod).

• Upload files to S3 via Presigned URLs.

• Supabase Auth + RLS on the client: understanding how roles and permissions work.

• Optionally (if a template is chosen): experience with SaaS templates (ability to remove unnecessary parts).

Stack:

• Next.js 14

• TypeScript

• React Query

• TailwindCSS + Shadcn UI

• React Hook Form + Zod

• Supabase Client (Auth + Realtime)

• Optionally: SupaStarter / ShipFast (if you choose this approach)

What is already available:

• HTML-Wireframes (layouts).

• Product Brief (UI/UX requirements).

AI Engineer (ComfyUI)

What needs to be done:

• Assemble and optimize a Docker image with ComfyUI and all models.

• Set up the Docker image for RunPod Serverless (ComfyUI + models + Python SDK).

• Optimize cold start (Network Volume Strategy: models on a network drive, not in the image).

• Implement callback processing (RunPod → webhook to backend).

• Set generation parameters (motion_bucket_id, fps, interpolation) to ensure smooth video.

• VRAM management (tiled_decode, lowvram) to prevent generation from crashing the worker.

• Content moderation: integration of trigger word filters for NSFW (checking prompts before generation).

Mandatory experience:

• ComfyUI: ability to assemble JSON workflows from nodes.

• Experience working with RunPod or similar (Modal, Replicate).

• Docker (building production images with CUDA support).

• Understanding VRAM limitations and optimization methods.

Stack:

• ComfyUI

• Python (for binding and integration with RunPod SDK)

• Docker

• RunPod Serverless

What is already available:

• Ready WorkFlows

• Critical User Flows (what generation should do).

How to respond?

Send:

1. Your role: (Backend Lead / Frontend / AI Engineer / Fullstack Lead).

2. Experience: Briefly about relevant projects (billing, asynchronous systems, ComfyUI, RunPod).

3. Timeline: How much time you will need to implement the MVP

4. Payment: FixPrice or Milestones (indicate your preferences) and your budget proposal considering the timeline.