\- No central coordinator
\- Nodes train locally on custom data shards
\- Aggregation (e.g., FedAvg) happens across verifiable nodes
\- All results are hash-verified before acceptance
\- Decentralized, docker-native FL infra
\- Ideal for research in Non-IID, private datasets, or public benchmark tasks
Project:
GitHub – [https://github.com/theblitlabs](https://github.com/theblitlabs)
Docs – [https://blitlabs.xyz/docs](https://blitlabs.xyz/docs)
We’re college devs building a trustless alternative to AWS Lambda for container-based compute, Federated learning and LLM inference
Would love feedback or help. Everything is open source and permissionless.