Enterprise-Grade Technology Stack
Built on proven blockchain infrastructure with novel consensus innovation
Solana Blockchain
65,000 TPS throughput
Sub-second finality
Filecoin/IPFS
Decentralized storage
Content-addressed data
Smart Contracts
Rust programs
On-chain verification
Encryption
End-to-end security
TEE support
System Architecture
AI Developers & Enterprises
Submit jobs via SDK/API
client.submit_job(model, data, gpu_type)
VFTChain Protocol
Job marketplace & consensus
Job Assignment
Validator Verification
Payment Settlement
GPU Mining Network
Process jobs, earn VFT
Proof of Useful AI Work (PoUAW)
Submit Job
Smart Contract
GPU Processing
Validators Verify
Result Delivered
Validator Network
Validator Role
Validators ensure the integrity of AI computations by verifying that miners have correctly executed jobs. They form a decentralized verification layer that prevents fraud and ensures quality.
Low Hardware Requirements
Any GPU or integrated graphics, 4GB+ RAM
Earn 5% of Job Fees
Validators share verification rewards
No Staking Required
Start validating immediately, no collateral needed
Verification Process
# Validator verification flow
1. Miner completes AI job
2. Result submitted to blockchain
3. Validator committee selected randomly
4. Validators run verification tests:
- Check intermediate checkpoints
- Verify gradient commitments
- Test on hidden validation data
5. 2/3 majority confirms result
6. Rewards distributed:
- 95% to miner
- 5% to validatorsEasy Integration
Python SDK Example
# Install VFTChain SDK
pip install vftchain
# Submit an AI training job
from vftchain import Client
client = Client(wallet="phantom_wallet_address")
# Upload dataset to IPFS/Filecoin
dataset_cid = client.upload_data("training_data.tar.gz")
# Submit GPU job
job = client.submit_job(
task_type="training",
model="transformer",
dataset=dataset_cid,
gpu_requirements={
"type": "RTX_4090",
"vram_gb": 24,
"count": 4
},
max_price_per_hour=2.50 # in VFT
)
# Monitor progress
print(f"Job ID: {job.id}")
print(f"Status: {job.status}")
print(f"Assigned to: {job.miner_address}")
# Get results
result = client.get_result(job.id)
model_weights = client.download(result.cid)Supported Frameworks
PyTorch - Full support for all versions
TensorFlow - Native integration
JAX - Accelerated computing
Hugging Face - Model hub support
Docker - Custom environments