AI and Web3 Research Lab

Build Smarter Onchain Systems

Access peer-reviewed AI models, Web3 primitives, and experimental toolkits designed for developers shipping at the intersection of machine learning and decentralized infrastructure.

47
Published Research Papers
12k
GitHub Stars
89
Open Source Repositories
6
Active Research Tracks
What we do

Decentralized AI research for the builder generation

Geltrano Labs operates as an independent research facility focused on the convergence of artificial intelligence and blockchain technology. We publish open-source implementations, conduct reproducible experiments, and release developer toolkits that solve real integration challenges. Our work spans neural network optimization for constrained environments, smart contract security analysis, and novel consensus mechanisms.

  • All research outputs ship with MIT-licensed reference implementations
  • Dedicated testnet environments for AI-chain integration experiments
  • Weekly research digests with implementation walkthroughs
  • Direct Discord access to researchers for technical deep-dives
Open Research
Reproducible science for everyone
Every experiment we conduct includes full methodology documentation and runnable code. We believe scientific progress in AI and Web3 requires radical transparency and community verification.
Builder-First Design
Tools that ship today
Our research outputs prioritize practical applicability over theoretical elegance alone. Each paper accompanies production-grade SDKs tested across multiple blockchain networks and inference environments.
Decentralized Collaboration
Global minds solving together
We operate as a distributed collective of researchers across twelve countries. This structure mirrors the systems we build and ensures diverse perspectives shape every project.
Process

From Research to Your Repository

01
Explore Active Tracks
Browse our current research initiatives spanning zero-knowledge ML, autonomous agents, and decentralized compute markets. Each track includes problem statements, progress updates, and contribution opportunities.
02
Clone Reference Code
Every published finding ships with a companion repository containing tested implementations. Fork directly into your project or use our package managers for cleaner dependency management.
03
Run Experiments Locally
Our sandbox environment replicates exact conditions from our papers. Modify parameters, swap models, and validate results against our published benchmarks on your own hardware.
04
Ship to Production
Graduate from experimentation to deployment using our hardened SDK releases. Production builds include monitoring hooks, upgrade paths, and dedicated support channels.

Start Building at the AI-Chain Frontier

Join thousands of developers using Geltrano research to ship the next generation of intelligent decentralized systems.

Features

Everything you need

Geltrano Labs operates as an independent research facility focused on the convergence of artificial intelligence and blockchain technology. We publish open-source implementations, conduct reproducible experiments, and release developer toolkits that solve real integration challenges. Our work spans neural network optimization for constrained environments, smart contract security analysis, and novel consensus mechanisms.

Neural Contract Auditor
Static analysis engine combining transformer models with formal verification for Solidity and Rust smart contracts. Catches vulnerability patterns that traditional tools miss by understanding semantic intent.
Onchain Inference Protocol
Framework for executing verifiable AI inference with cryptographic proofs anchored to any EVM chain. Enables trustless machine learning predictions in decentralized applications.
Model Compression Toolkit
Quantization and pruning library optimized for resource-constrained blockchain nodes. Reduces model sizes by up to ninety percent while preserving accuracy for edge deployment.
Distributed Training Harness
Coordinate model training across heterogeneous compute providers with built-in incentive alignment. Supports federated learning scenarios where data cannot leave local environments.
Chain Data Pipeline
ETL framework specifically designed for blockchain data feeding into ML workflows. Handles reorgs, mempool analysis, and cross-chain state aggregation automatically.
Research Sandbox
Pre-configured development environment with our complete toolkit stack. Spin up isolated experiments in minutes with deterministic reproducibility guarantees.
How it works

From Research to Your Repository

We approach research as builders rather than pure academics. Every project begins with a real developer problem sourced from community forums, hackathon feedback, and protocol team consultations. We then work backward to identify which scientific questions need answers. This inverted methodology keeps our output grounded in genuine utility.

01
Explore Active Tracks
Browse our current research initiatives spanning zero-knowledge ML, autonomous agents, and decentralized compute markets. Each track includes problem statements, progress updates, and contribution opportunities.
02
Clone Reference Code
Every published finding ships with a companion repository containing tested implementations. Fork directly into your project or use our package managers for cleaner dependency management.
03
Run Experiments Locally
Our sandbox environment replicates exact conditions from our papers. Modify parameters, swap models, and validate results against our published benchmarks on your own hardware.
04
Ship to Production
Graduate from experimentation to deployment using our hardened SDK releases. Production builds include monitoring hooks, upgrade paths, and dedicated support channels.
FAQ

Common questions

How do I access your research outputs?+
All papers and code publish to our GitHub organization under MIT license. No signup required for public repositories. Premium research tracks with early access require a developer membership.
Can I contribute to ongoing research?+
Absolutely. Each active research track maintains a public roadmap with tagged contribution opportunities. Submit proposals through our governance forum or jump directly into open issues.
What blockchain networks do your tools support?+
Our core libraries target EVM-compatible chains with dedicated modules for Ethereum, Polygon, and Arbitrum. Solana and Cosmos SDK support ships in our next major release cycle.
Do you offer enterprise partnerships?+
We collaborate with protocol teams and enterprises on sponsored research tracks. These partnerships fund our open research while giving partners priority input on problem selection.
How is Geltrano Labs funded?+
We sustain operations through a mix of grants from major ecosystem foundations, enterprise research partnerships, and a small developer membership program for premium content access.
About us

The Story Behind the Lab

Geltrano Labs advances open AI and blockchain research to equip developers with production-ready tools that bridge intelligent systems and decentralized networks.

Geltrano Labs started in late 2022 when four researchers grew frustrated watching promising AI-blockchain integration papers gather dust without implementations. They pooled savings, converted a Lisbon warehouse into a compute cluster, and committed to publishing working code alongside every theoretical contribution. That founding principle remains unchanged.

We approach research as builders rather than pure academics. Every project begins with a real developer problem sourced from community forums, hackathon feedback, and protocol team consultations. We then work backward to identify which scientific questions need answers. This inverted methodology keeps our output grounded in genuine utility.

The future belongs to systems that think onchain and to the developers brave enough to build them today.

Looking ahead, we see AI and decentralized systems becoming inseparable infrastructure layers. Our vision positions Geltrano Labs as the essential bridge connecting these domains through rigorous research and accessible tooling. We measure success not by citation counts but by production deployments running code we helped create.

Values

What we stand for

Radical Openness
We publish methodologies, failures, and financial operations alongside successes. Transparency builds trust and accelerates collective progress faster than competitive secrecy ever could.
Rigorous Pragmatism
Ideas must survive contact with compilers and testnets. We maintain high academic standards while ruthlessly prioritizing research that developers can actually ship.
Patient Ambition
Fundamental breakthroughs require sustained effort over years, not quarters. We structure our funding and timelines to support deep work without chasing short-term trends.
Team

The people behind Geltrano Labs

MV
Marta Voskovic
Research Director
Former DeepMind researcher specializing in reinforcement learning and mechanism design. Led the team that developed our onchain inference verification protocol.
JO
James Okonkwo
Protocol Architect
Built core infrastructure at two major L1 blockchains before joining Geltrano. Obsessed with making distributed systems feel simple to application developers.
LZ
Lena Zhang
ML Engineering Lead
Spent six years optimizing production ML systems at scale before pivoting to Web3. Champions our model compression research and edge deployment initiatives.
RA
Ravi Anand
Developer Relations
Former hackathon organizer turned researcher advocate. Ensures every paper we publish comes with documentation that actually helps builders implement our findings.
Our mission

Decentralized AI research for the builder generation

Contact

Get in touch

We respond to all messages within 24 hours.

Response time
Within 24 hours
Location
Remote — Global

Find us online

Twitter / X GitHub LinkedIn Email

Stay updated

Get updates on new features and research.