Loading
AI Series
COMPUTE AT THE FRONTIER.
GPU-accelerated workstations and multi-GPU systems for machine learning training, inference, data science, and AI research. Built for the computational demands of modern AI.
Starting from₦5,000,000
WHO IT'S FOR
AI ResearchersML EngineersData ScientistsResearch LabsUniversitiesAI StartupsNLP Engineers
OPTIMIZED FOR
PyTorchTensorFlowCUDAJupyterPythonDockerKubernetesHugging FaceJAXRAPIDS
What Our Clients Say
BUILT FOR PROFESSIONALS LIKE YOU
Training times on my computer vision models dropped from overnight runs to under two hours. The RTX 4090 with 24GB VRAM handles the batch sizes I actually need — I haven't had to compromise on model architecture since switching to the AI Series Professional.
Ibrahim M.
ML Engineer
Abuja, Nigeria
We procured two Lab tier workstations for our research group. Sephora handled the institutional purchase process professionally, provided a formal quote for our finance department, and delivered pre-configured with Ubuntu, CUDA, and PyTorch. Exactly what we needed.
Dr. Aisha K.
AI Researcher, University of Abuja
Abuja, Nigeria
Running LLM fine-tuning locally instead of paying cloud GPU bills every month made the AI Series pay for itself in about four months. The Sephora team knew exactly how to configure the CUDA stack — zero setup headaches.
Seun B.
Data Scientist & Startup Founder
Lagos, Nigeria
Configurations
RECOMMENDED TIERS
Each tier is a starting point. Every system can be fully customized to your workflow.
RESEARCH
For model experimentation, fine-tuning, and small-scale training
- Intel Core i7-14700K / AMD Ryzen 9 7900X
- 64GB DDR5 RAM
- NVIDIA RTX 4070 Ti Super 16GB
- 2TB NVMe SSD
- 850W 80+ Gold PSU
Starting from
₦5,000,000
70% deposit to confirm · 30% balance before delivery
RECOMMENDED
PROFESSIONAL
For serious training workloads and multi-model inference
- Intel Core i9-14900K / AMD Ryzen 9 7950X
- 128GB DDR5 RAM
- NVIDIA RTX 4090 24GB
- 4TB NVMe SSD
- 1200W 80+ Platinum PSU
- Optimized CUDA and driver stack
Starting from
₦6,200,000
70% deposit to confirm · 30% balance before delivery
LAB
For large-scale training, multi-GPU, and production inference
- AMD Threadripper / Intel Xeon W
- 256GB DDR5 ECC RAM
- 2x NVIDIA RTX 4090 24GB (or A6000)
- 8TB NVMe SSD RAID
- 1600W Redundant PSU
- Server-grade chassis with enhanced cooling
Starting from
₦12,000,000+
70% deposit to confirm · 30% balance before delivery
The Process
HOW IT WORKS
01
CONFIGURE
Use our online configurator or speak with our team to spec your exact build.
02
CONFIRM & DEPOSIT
We lock in your price. A 70% deposit confirms your order and starts procurement.
03
BUILD & TEST
We assemble, stress-test, and quality-check your system. Typically 5–7 working days.
04
DELIVER & SETUP
Your system is delivered and set up. 30% balance is due before delivery. 1-year warranty starts on delivery day.
FREQUENTLY ASKED QUESTIONS
What GPU is best for machine learning?
For most ML workloads, the RTX 4090 with 24GB VRAM offers the best price-to-performance ratio. For production and very large models, NVIDIA A6000 (48GB) or multi-GPU configurations are recommended.
How much VRAM do I need?
It depends on model size. For fine-tuning LLMs and training medium-sized models, 24GB (RTX 4090) is excellent. For very large models, multi-GPU setups or 48GB cards may be necessary.
Can I add more GPUs later?
Our Professional and Lab tiers are designed with multi-GPU expansion in mind. The motherboard, PSU, and chassis are selected to support additional GPUs.
Do you pre-install CUDA and ML frameworks?
Yes. We can pre-configure your system with Ubuntu, CUDA toolkit, cuDNN, PyTorch, TensorFlow, Docker, and any other tools your workflow requires.
Do you support university and research lab purchases?
Yes. We handle institutional procurement, provide formal quotes, and offer service agreements for research installations. Visit our Enterprise page for details.
What about NVLink for multi-GPU?
NVLink support depends on the specific GPU and motherboard configuration. Our team can advise on the best multi-GPU setup for your training requirements and budget.
READY TO CONFIGURE YOUR AI BUILD?
Use our configurator to spec out your perfect system, or speak directly with our team for tailored advice.
Sephora Systems AI Series workstations are purpose-built for artificial intelligence and machine learning professionals in Nigeria and Africa. Whether you're training deep learning models with PyTorch and TensorFlow, running inference workloads, processing large datasets with RAPIDS, or developing NLP applications with Hugging Face, our GPU-accelerated systems provide the compute power you need. From single-GPU research workstations to multi-GPU lab systems, the AI Series delivers serious computational capability with local support, Naira pricing, and expert consultation from our Abuja team.