Origin Story
ResonTech started at ETHKyiv 2025. Sven, Kyrylo, Petro, and Ivan built a prototype decentralized GPU marketplace on Ethereum — idle GPUs as liquid compute with state channel settlement. Won the DeAI and State Channels tracks.
But we learned something unexpected: the hardest part wasn't the payment layer. It was the orchestration. Getting distributed jobs to actually run — reliably, fast, automatically — was unsolved for most ML teams.
We pivoted. Dropped the blockchain rails. Rebuilt from scratch as a proper ML infrastructure platform. Dmytro joined and the vision grew: not just a GPU marketplace, but an end-to-end ML ecosystem covering every stage of the lifecycle.
| Date | Milestone |
|---|---|
| Summer 2025 | ETHKyiv 2025 — won DeAI and State Channels tracks with a decentralized GPU marketplace prototype. |
| Summer 2025 | Pivoted from blockchain to proper ML infrastructure. Dmytro Horskyi joined as Head of Infrastructure. |
| Fall 2025 | Built the core kernel, public GPU pool, per-account S3 bucket (Garage) with presigned-URL data access, and pay-as-you-go billing. Python SDK and web platform shipped. |
| Winter 2025 | Public pool opened to beta testers. First real training jobs — DeepLab and image model fine-tuning. Zero lost checkpoints. |
| 2026 | Beta testing managed clusters and inference endpoints. Dedicated nodes, isolated kernel, first live inference runs. |
Mission
AI teams waste enormous time managing infrastructure instead of doing science. Configuring Kubernetes clusters. Debugging NCCL errors. Recovering from spot instance preemptions. Reconciling GPU billing. Hours that should be spent on model architecture and iteration.
Our mission is to eliminate that entirely. Submit a job. Get your model back. That's it.
Values
Zero Infrastructure Headaches
Researchers should spend time on models, not kubectl. Every abstraction removes another operational burden.
Radical Transparency
Real-time telemetry. Live job logs. Honest pricing. No opaque billing or invisible failures.
Privacy by Architecture
Gradient-only federated learning means raw training data never leaves your infrastructure. Privacy isn't a feature — it's a design constraint.
Global Access
World-class GPU infrastructure should not be limited to teams with AWS enterprise agreements. We're building for every ML team, everywhere.
The Team
Contact
| Topic | Contact |
|---|---|
| General enquiries | office@reson.tech |
| Enterprise & Private Cluster | office@reson.tech |
| GPU supplier onboarding | office@reson.tech |
| Press & partnerships | office@reson.tech |