Mayke De Freitas Santos

This page shows what I am working on at the moment.

GitHub Email
HardwareHPC

Single-Tower High-Performance Rig

While most people drop thousands of £GBP on a depreciating Audi Q3 that loses value the moment they drive it off the lot (or worse, they lease it), I'd rather sink that money into something that can actually produce interesting things. This machine currently runs 70B+ parameter models locally and processes terabyte-scale microscopy datasets. The full build-out to three GPUs and 1TB of RAM will generate value every day and stay relevant as I upgrade components.

The core philosophy here is modularity without compromise. The Threadripper 7970X gives me 48 PCIe 5.0 lanes, meaning I can run three flagship GPUs at full x16 bandwidth without needing £GBP thousands more on a dual-socket EPYC or enterprise Xeon platform. The ASUS Pro WS TRX50-SAGE has IPMI and ECC support, so this isn't just a gaming rig with delusions of grandeur—it's production-grade infrastructure.

This started as a microscopy research platform. I needed something that could crunch through days of high-resolution imaging data while simultaneously running LLM inference for automated experimental workflows. Turns out, the Venn diagram of "can process microscopy data" and "can run massive language models" is just "absurdly powerful computer."

ECC memory isn't negotiable. When you're running multi-day experiments or training on scientific data, a single bit flip can corrupt everything. The 128GB DDR5 ECC setup expands to 1TB because why set artificial limits?

The 96TB ZFS pool means I can stop worrying about cloud storage costs and actually work with real datasets locally. Checksumming, snapshots, data integrity—all the things you want when your data represents months of work.

The RTX 6000 Blackwell (96GB VRAM) is cutting-edge enough to stay relevant as models get bigger, and when I need more, I'll just add two more GPUs. 288GB total VRAM in a single tower. No racks, no datacenter, no monthly AWS bills making me cry.

Sub-systemComponentRationale
CPUAMD Threadripper 7970X32 cores, 48 PCIe 5.0 lanes.
GPUNVIDIA RTX PRO 6000 96GBBlackwell architecture for large-scale compute.
RAM128 GB DDR5-4800 ECCRDIMM; Expandable up to 1TB.
NVMe OSCrucial P2 1TBDedicated OS drive.
NVMe ScratchSamsung 9100 Pro 8TB8TB scratch storage for active datasets.
Bulk Storage96TB ZFS Pool4x Seagate IronWolf Pro 24TB drives.
MotherboardASUS Pro WS TRX50-SAGEIPMI and ECC support.
SoftwareVision-RAG

ColPali Research Engine

A containerized Vision-RAG pipeline that bypasses traditional OCR to perform visual document retrieval directly on document images.

# Blackwell-Optimized Deployment
./manage_env.sh
StackTechnologyImplementation
EnvironmentCUDA 12.6.2PyTorch Nightly for sm_120 support.
RetrieverColPali v1.2200 DPI multi-vector indexing.
ReaderQwen2-VL-7BVisual reasoning with grounding enforcement.
StorageNVMe Volume MountDirect access to 8TB scratch NVMe.