ESTABLISHMENT OF ARTIFICIAL INTELLIGENCE (AI) TRAINING AND INNOVATION LAB AT HINDU COLLEGE AS PER QCBS METHODOLOGY

Amount

Supply, Installation, Testing, and Commissioning of High-Performance Workstation with the following specifications: • CPU: Intel Core i9 14th Gen or higher • RAM: 64GB DDR5 5600 MHz, expandable • GPU: NVIDIA RTX 2000 Ada Generation (16GB VRAM) or better • Storage: 2TB PCIe NVMe M.2 SSD (Gen 4 or higher) • Display: 27" Professional Ultra HD Monitors (4K, IPS, HDR) • OS: Windows 11 Pro (Licensed) with Ubuntu LTS via dual-boot or virtual machine • Software Features: Remote Graphics Collaboration HD • Keyboard: AI-Optimized (Ergonomic, Programmable Macro Keys) • Power Supply: Min. 700W Platinum Rated, GPU-Ready or higher • Warranty: 3 Years OEM Smart Pack with On-Site Support Software (Pre-installed, Configured and Licensed): OS and IDE licenses including Win 11 Pro, Ubuntu LTS, and PyCharm Professional (1-year subscription with advanced AI/ML, scientific computing, and web development support). All licenses to be provided digitally with activation keys and compliance documentation. AI/ML/DL stack licenses and libraries including TensorFlow, PyTorch, Hugging Face, Scikit-learn, and LangChain. Includes enterprise or commercial licenses (where applicable), access to official repositories, API integrations, and premium support subscriptions if required. All licenses to be provided digitally with activation details and compliance documentation.

A centralized, remote management solution shall be provided to enable secure, out-of-band access and diagnostics for all 30 workstations. The system must allow administrators to remotely monitor system health, perform BIOS-level configurations, push firmware updates, and execute power control (on/off/reset) across all workstations from a single interface. It must also support full KVM (keyboard, video, mouse) redirection over IP for each connected workstation, ensuring complete access and control even during pre-boot phases or operating system failure. The solution must function independently of the host OS and provide a browser-based interface for ease of access and administration.

Supply, Installation, Testing, and Commissioning of High-Performance Server with the following minimum specifications (or better): Chassis & Power System: Enterprise-grade HPE ProLiant DL380a Gen12 (or equivalent) server chassis suitable for continuous AI/ML workloads, equipped with 8 × Hot-Swappable 2400W Titanium Redundant Power Supply Modules (M-CRPS) and optimized thermal design supporting multi-GPU configurations. Processors: 2 × Intel Xeon 6747P CPUs, each with 48 cores / 96 threads, 2.7 GHz base frequency, 5th Gen Intel Xeon Scalable architecture, 330W TDP, with Intel AMX acceleration. Memory: 512 GB DDR5 ECC Registered Memory, configured as 8 × 64GB DDR5-6400 RDIMMs, expandable as per OEM maximum capability. Storage Subsystem: 1 × 1.92 TB NVMe Gen4 U.3 SSD for operating system/boot, 4 × 3.84 TB NVMe Gen4 U.3 SSDs for data (total raw ~17.28 TB), Read-intensive enterprise-class NVMe SSDs. Storage Controller: HPE MR416i-p Gen11 Hardware RAID Controller with support for RAID 0/1/5/6/10 across NVMe/SAS/SATA drives, 8GB cache with power-loss protection, and HPE 96W Smart Storage Battery. Networking: Dual-port 10GbE OCP3 Ethernet Adapter with 2 × 10Gb SFP+ SR Transceivers, along with dedicated 1GbE management port (IPMI/iLO) supporting secure remote configuration and monitoring. GPU Accelerators: 4 × NVIDIA H200 NVL Tensor Core GPUs, each with 141GB HBM3 memory, suitable for large-scale AI/ML model training, deep learning, generative AI, and HPC workloads. Includes 2 × NVIDIA NVLink Bridges for high-bandwidth GPU-to-GPU interconnect. Management & Monitoring: Integrated HPE iLO Advanced remote management (3-year license) with remote KVM, virtual media, telemetry, diagnostics, and secure firmware management. Includes HPE Compute Ops Management SaaS (5-year license). Expansion & Accessories: PCIe Gen5 architecture with minimum 1 × PCIe Gen5 x16 slot available for future expansion. Includes all OEM-recommended GPU power kits, NVMe–Tri-Mode PCIe enablement kits, OCPA cabling, and ball-bearing rail kit for rack installation. Software Stack: Pre-installed and validated Ubuntu Linux LTS, NVIDIA Base Command OS components, CUDA Toolkit, cuDNN, NCCL, TensorFlow, PyTorch, Docker CE with NVIDIA Container Runtime, fully optimized for GPU compute workloads. Warranty & Support: 5 Years OEM Enterprise Support including software updates, remote diagnostics, and RMA service.

Swan Solutions and Services Pvt Ltd (BID ID -3332208)

Sanbay Networks Pvt Ltd (BID ID -3332448)