Contact person: Business Client

Phone:Show

Email:Show

Location: DUBAI, United Arab Emirates

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"We are seeking an experienced Software Development Agency or Freelance Team to design, develop, and deliver a multi-tenant AI based SaaS platform with integrated AI/ML (LoRA-based NLP), sponsor & enterprise dashboards, and admin controls. The platform will serve SMEs and enterprise users with tiered product offerings, analytics, and payment integrations — built on a scalable, modular architecture.

Core System Development
• Multi-tenant SaaS Architecture with separate accounts for SMEs, Sponsors, and Admin.
• Dynamic Feature Toggling system – Admin can enable/disable modules per tier or tenant.
• Admin Dashboard: Tenant management, billing, analytics, compliance controls.
• Sponsor Dashboard:
◦ License allocation and usage monitoring
◦ Token pack purchasing & payment integration
◦ SROI & usage analytics
◦ Token consumption forecasting & auto-renewal options
• SME Dashboard: Chatbot setup, analytics, onboarding wizard.

AI/ML Infrastructure
• Language Model: Google Gemma 3 (4B) with LoRA adapters and other required.
• Sector-specific LoRA Adapters: Retail, Healthcare, Finance, etc.
• Dynamic Runtime Linking of LoRA adapters per tenant.
• LoRA Retraining Framework: Ability to retrain models using new datasets.
• Multi-Tenant Isolation: Manage concurrent LoRA adapters using S-LoRA with caching.
• Precision & Quantization: BF16 (Bfloat16).

Web & Dashboard Development
• Frontend: Public marketing website + SaaS dashboards.
◦ Homepage, Product Tiers, Pricing, Sponsors, Case Studies, Contact.
• Private Dashboards:
◦ SME Portal, Sponsor Dashboard, Master Admin Dashboard.
• Technology Stack: React.js / [login to view URL], TailwindCSS, TypeScript preferred.
• Design: Responsive, modern, clean UI/UX (Figma or similar).

Backend & Integrations
• API-Driven Architecture (REST / GraphQL).
• Backend Stack: Node.js / Python / Go (open to recommendation).
• Database: PostgreSQL / MongoDB (multi-tenant support).
• Integrations:
◦ Payment: SnapScan, Zapper, Stripe.
◦ Communication: WhatsApp API, Email/SMS gateway.
◦ Auth: OAuth 2.0 / JWT.
• Hosting: AWS Cape Town (POPIA-compliant).
• Analytics: Token usage analytics via API integration.

Enterprise & Add-On Modules
• Advanced Analytics Package
• AI Model Management Suite (LoRA retraining controls)
• Enterprise Integration Suite (CRM/ERP connectors)
• Sector-specific LoRA development
• Deep usage and staff adoption analytics

Technical & Security Requirements
• POPIA Compliance: Consent, data retention, audit logging.
• Scalability: Auto-scaling & load balancing for dynamic workloads.
• Security: Role-based access control, encrypted data storage, secure payments.
• Deployment: CI/CD pipeline with staging + production environments.

Required Skills & Expertise
• SaaS Product Development
• Multi-tenant Architecture
• React.js / [login to view URL] / TailwindCSS
• Node.js / Python / FastAPI / Go
• PostgreSQL / MongoDB
• AI/ML Model Integration (LoRA / Hugging Face / Gemma 3)
• API Development & Integration
• AWS Cloud (EC2, S3, IAM, CloudWatch, GPU setup)
• Figma / UI/UX Design
• Payment Gateway Integration
• Data Security & Compliance (GDPR / POPIA)

Tools & Technologies
Frontend: React, [login to view URL], Tailwind, Redux
Backend: Node.js, FastAPI, Django, or [login to view URL]
Database: PostgreSQL / MongoDB
ML: Google Gemma 3, PyTorch, LoRA, S-LoRA
DevOps: AWS, Docker, CI/CD (GitHub Actions / Jenkins)
Integrations: Stripe, WhatsApp API, SnapScan, Zapper
Analytics: Custom dashboard + API-based token analytics

Engagement Type
• Fixed-cost or milestone-based engagement.
• End-to-end delivery or module-wise collaboration accepted.
• Long-term partnership potential for global rollout and ongoing maintenance.

Ideal Partner Profile
• Proven experience in building AI & ML, SaaS or enterprise-grade web applications.
• Strong communication and agile project management capability.
• Ability to work with distributed teams across time zones." (client-provided description)