J
Jarvis
Agent runtime that spawns, configures, and secures Personal Operators.
Products
A system, not a collection
Flow Research's products form one system: Jarvis gives the agent life, Garden gives it a workspace, WorkStream gives it valuable work, and Harnessy makes it reliable. Contributors help build every layer.
G
Garden
Human-agent workspace with connected tools, workflows, and approvals.
W
WorkStream
Task pipeline that distributes work, verifies output, and handles rewards.
H
Harnessy
Reliability layer for testing, evaluating, and improving agent behavior.
Curriculum
A clear path from fundamentals to contribution
The curriculum is organized around the capabilities you need to contribute to building Flow Research's products — from foundations through production-ready systems.
9 lessons
Foundations
Engineering fluency
Reading, documentation, version control, collaboration, and the mindset needed to work in public.
18 lessons
Blockchain
Decentralized infrastructure
Consensus, smart contracts, security, scalability, and protocol economics from beginner to advanced.
18 lessons
AI/ML
Production-minded machine learning
Math, pipelines, lifecycle, notebooks, libraries, MLOps, architectures, and applied research practice.
18 lessons
Protocol Engineering
Systems that coordinate
State machines, specifications, resilience, governance, performance, and enterprise-grade adoption.
Learning model
Built for builders who learn by making things real
Flow Research is not a content library for passive reading. Each area is designed to help learners turn concepts into notes, code, diagrams, experiments, and public contributions.
Browse the full curriculum- Read technical material like a builder, not a passive student.
- Move from concept notes into reproducible labs and visible artifacts.
- Explain tradeoffs clearly across AI, blockchain, and distributed systems.
- Build toward open-source contribution instead of isolated coursework.
- Earn points and reputation through verified public contributions.
Launch-ready learning