RMD Solutions

National Shared
AI Infrastructure

An interactive walkthrough of the user journey — from login to innovation on the platform.

IaaS PaaS AI/ML Pay-Per-Use

↓ Use arrow keys to navigate • Press 'F' for Full Screen • Interactive demos on each slide

The Challenge

Why Institutions Struggle Today

Key challenges faced by Skill Universities and ITIs in establishing independent AI infrastructure.

💰

High Upfront CapEx

High upfront capital expenditure for GPUs, servers, and storage.

⏳

Rapid Obsolescence

Rapid obsolescence of GPUs and computing infrastructure.

👨‍💻

Talent Shortage

Shortage of specialised AI, Cloud, and DevOps administrators.

🌙

Under-utilisation

Under-utilisation of expensive GPU infrastructure.

📑

Duplicated Procurement

Multiple institutions purchasing similar infrastructure independently.

🔒

Security & Licensing

Complex software licensing, maintenance, and cybersecurity challenges.

The Solution

Proposed Concept — AI Infrastructure as a Utility

A shared, catalogue-driven, pay-per-use national platform — so institutions focus entirely on learning and innovation.

"I need 1 GPU, 16 vCPU, 64 GB RAM, Python, Jupyter, PostgreSQL and an LLM environment for 48 hours."

A typical request. The platform provisions it automatically.
Personas

Meet the Users

The platform serves the entire Skill Development ecosystem.

Student

Priya
Student

Capstone project on predictive maintenance using AI/ML.

Faculty

Prof. Sharma
Faculty

120 students in 3 batches.
GPU environments for deep learning practicals.

Researcher

Dr. Meena
Researcher

NLP researcher working on fine-tuning language models on Indian languages.

1Login
2Browse
3Configure
4Provision
5Build
6Monitor
7Pay

Portal Login

Users access the self-service portal with institutional SSO.

🔐 ACCESS CONTROL
Role-Based Access
Student • Faculty • Researcher • Admin
Resource Quotas
GPU-hours • Storage • API calls per role
Institutional SSO
SAML / OAuth • Existing university credentials
🔒 ai-lab.skill.gov.in/login
✓Login
2Browse
3Configure
4Provision
5Build
6Monitor
7Pay

Self-Service Catalogue

Pre-configured AI models, compute profiles, and developer stacks ready for instant deployment.

🔒 ai-lab.skill.gov.in/catalogue
👤 Priya Kumar • Student
All Services
Compute
AI / ML
Data & Storage
Dev Tools
🖥️
CPU / vCPU
General VMs
⚡
GPU
A100 / H100
📦
Containers
Docker / K8s
💾
Storage
Object / Block
🗄️
SQL DB
PostgreSQL
🔮
Vector DB
Embeddings
🧠
AI/ML
PyTorch / TF
🤖
GenAI / LLM
Models & RAG
🔗
Agentic AI
Agent Framework
📊
Analytics
BI & Dashboards
📓
Jupyter
Notebooks
🛠️
Dev Tools
IDE / Git / APIs
0 services selected
✓Login
✓Browse
3Configure
4Provision
5Build
6Monitor
7Pay

Configure Your Environment

Drag the sliders to adjust resources. Cost updates in real time.

🔒 ai-lab.skill.gov.in/configure
⚙️ RESOURCE CONFIGURATION
⚡ GPU (A100)
1
🖥️ vCPU Cores
16
💾 RAM (GB)
64 GB
📀 Storage (GB)
100 GB
⏱️ Duration (hours)
48 hrs
Estimated Cost*
₹ 4,320
💳 COST BREAKDOWN
⚡ GPU (A100) ₹ 3,360
🖥️ vCPU Cores ₹ 384
💾 RAM Memory ₹ 384
📀 Storage Volume ₹ 150
Total* ₹ 4,320
📦 SOFTWARE & PLATFORM STACK
Python 3.11 Jupyter PyTorch 2.3 PostgreSQL CUDA 12 LLM API
* Indicative figures for representation purposes; subject to final commercial agreements.
✓Login
✓Browse
✓Configure
4Provision
5Build
6Monitor
7Pay

Auto-Provisioning

Zero-touch environment orchestration and instant cloud resource allocation.

🔒 ai-lab.skill.gov.in/env/env-2847
🚀 DEPLOYMENT STATUS
  • ⏳ Allocating GPU resources (1× A100)
  • ⏳ Provisioning vCPU (16 cores) & RAM (64 GB)
  • ⏳ Attaching storage volume (100 GB SSD)
  • ⏳ Installing OS & CUDA drivers
  • ⏳ Deploying Python, Jupyter, PyTorch
  • ⏳ Starting PostgreSQL database
  • ⏳ Configuring LLM API access
  • ⏳ Running security checks
  • ⏳ Environment ready!
🏗️ WHAT'S HAPPENING

The platform automatically orchestrates across the three-layer architecture:

Layer 1 — Infrastructure
GPU + CPU + Storage + Network allocated
Layer 2 — Platform
OS + Containers + DB + ML Frameworks deployed
Layer 3 — Learning
Jupyter + LLM API + Project workspace ready
⏱️
Typical Provisioning Time: ~3–5 minutes
vs. weeks or months for traditional hardware procurement
⚡ Instant Allocation
✓Login
✓Browse
✓Configure
✓Provision
5Build
6Monitor
7Pay

Build & Learn

Integrated Jupyter notebook workspace pre-loaded with PyTorch, CUDA, and datasets.

🔒 ai-lab.skill.gov.in/workspace/env-2847/jupyter
🟢 GPU Active • 1× A100
📓 predictive_maintenance.ipynb
▶ Run
⏹ Stop
🔄 Restart Kernel
Python 3.11 | PyTorch 2.3 | CUDA 12
In [1]▶
# Load sensor dataset for predictive maintenance import torch import pandas as pd from torch import nn df = pd.read_csv('sensor_data.csv') print(f"Dataset: {df.shape[0]} samples, {df.shape[1]} features") print(f"GPU: {torch.cuda.get_device_name(0)}")
In [2]▶
# Build LSTM model for failure prediction class PredictiveModel(nn.Module): def __init__(self): super().__init__() self.lstm = nn.LSTM(18, 128, batch_first=True) self.fc = nn.Linear(128, 1) model = PredictiveModel().cuda() print(f"Model params: {sum(p.numel() for p in model.parameters()):,}")
In [3]▶
# Train and evaluate train_model(model, train_loader, epochs=50) accuracy = evaluate(model, test_loader) print(f"✅ Test Accuracy: {accuracy:.1%}") print(f"🏆 Model ready for deployment!")
✓Login
✓Browse
✓Configure
✓Provision
✓Build
6Monitor
7Pay

Usage Dashboard

Real-time visibility into infrastructure utilization, active sessions, and resource consumption.

🔒 ai-lab.skill.gov.in/dashboard
2,418
Active Users
15
Institutions
842
Environments
₹4.2L
This Month
GPU Utilisation0%
CPU Utilisation0%
Storage Used0%
🕐 RECENT ACTIVITY
  • Priya Kumar provisioned env-2847 2m ago
  • Prof. Sharma created 50 student envs 15m ago
  • Dr. Meena requested 4-GPU cluster 1h ago
  • ITI Chennai onboarded 200 students 3h ago
  • Auto-scaled GPU pool +8 units 4h ago
📈 TOP SERVICES
GPU Compute68%
Jupyter Notebooks54%
GenAI / LLM APIs41%
PostgreSQL33%
✓Login
✓Browse
✓Configure
✓Provision
✓Build
✓Monitor
7Pay

Consumption Billing

Transparent, itemised usage billing aggregated at the institutional level.

🔒 ai-lab.skill.gov.in/billing
📄 Invoice #INV-2847
Session:
48 hours
ResourceUsageRate*Amount*
GPU (A100)1 × 48 hrs₹70/GPU-hr₹ 3,360
vCPU16 × 48 hrs₹0.50/vCPU-hr₹ 384
RAM64 GB × 48 hrs₹0.125/GB-hr₹ 384
Storage100 GB × 2 days₹2.25/GB-mo₹ 15
PostgreSQL1 × 48 hrs₹3/inst-hr₹ 144
LLM API2,340 calls₹0.02/call₹ 47
Total Payable*₹ 4,334
PRICING MODEL*
GPU
₹/GPU-hour
CPU
₹/vCPU-hour
Storage
₹/GB-month
Database
₹/instance-hour
AI Model
₹/inference/API usage
Training Env
₹/student/month
ESTIMATED COST PER LEARNER / HOUR*
₹ 42
⚠️ vs. ₹Cr+ for independent infrastructure
* Indicative figures for representation purposes; subject to final commercial agreements.
Architecture

Three-Layer Architecture

Architecture

Learning & Innovation

AI Labs • GenAI • Agentic AI • Projects

Platform

OS • Containers • Databases • ETL • MLOps • APIs

Infrastructure

Compute • GPU • Storage • Network • Security

Lifecycle

Platform Lifecycle Management

Full-stack platform management by RMD Solutions — so institutions focus entirely on learning & innovation.

01 Plan
02 Provision
03 Onboard
04 Train
05 Operate
06 Monitor
07 Support
08 Upgrade
09 Optimise
10 Retire
🏛️

For Government

  • Lower CapEx via shared infrastructure pool
  • Centralized governance & security oversight
  • Higher GPU & compute asset utilization
  • Transparent pay-per-use consumption metrics
🏫

For Institutions

  • Zero upfront hardware investment or maintenance
  • Rapid environment provisioning in minutes
  • Always current GPU, software, and AI stack
  • Zero DevOps or cloud admin burden
👩‍🎓

For Students

  • Access to industry-grade A100 GPU compute
  • Hands-on learning with Jupyter & PyTorch
  • Real-world project development & AI modeling
  • Production-ready development environments
Roadmap

Proposed Rollout

Phase I — Pilot

5–10 institutions
Regional diversity
Validate the model

Phase II — Expansion

Expanded catalogue
More institutions
Based on outcomes

Phase III — National

Full ecosystem
Shared AI utility
Pan-India coverage

Impact

National Impact

🌍

Democratisation

AI beyond premier institutions

🛠️

Hands-on Education

Build, not just study AI

👨‍🏫

Faculty Transform

AI-enabled curriculum tools

🏭

Industry Ready

Real-world tech exposure

🔬

Research Enabled

On-demand HPC, no delays

📐

Standardisation

Common AI envs across India

The Vision

"AI Infrastructure
for Every Skill Learner"

Learn AI → Build AI → Experiment → Deploy
Create AI-enabled Solutions

RMD Solutions Private Limited
Programme Technology & Lifecycle Management Partner