Co-founder / Architect - Decentralized AI Platform - Equity only

P2P Decentralized AI Network Architect & Co-Founder
UBIQS | Founding role | Equity only at this stage
UBIQS is developing a trusted, peer-to-peer AI network that can coordinate AI workloads across data-center, cloud, and edge systems. We’re looking for a hands-on technical co-founder to design and build its networking and distributed-computing architecture.
This role spans the full stack: peer-to-peer protocols, distributed AI workloads, consensus and verification, and accelerator and chip-level co-design. It is a founding operator role, not an advisory position.
What you’ll own
Design and implement P2P networking protocols in Rust for secure, high-performance communication among decentralized AI nodes.
Build peer discovery, connectivity, routing, data exchange, and coordination across diverse cloud and edge environments, using technologies such as TCP/IP, UDP, WebRTC, and libp2p.
Design distributed AI execution, including model sharding and parallelism, federated learning, workload placement, and coordination across nodes.
Develop distributed inference architectures, including prefill/decode disaggregation, KV-cache transfer and placement, scheduling, and routing.
Work with blockchain engineers and protocol developers to integrate consensus, smart contracts, and cryptographic mechanisms for network coordination and settlement.
Evaluate proof mechanisms—including Proof of Work, proof of useful computation, and zkML—for verifying work or results, and understand their security, performance, and cost trade-offs.
Build for fault tolerance and resilience: handle node churn and failure, unreliable links, stragglers, synchronization, checkpointing, recovery, and untrusted participants.
Co-design AI workloads with accelerators and chips. Help shape compute, memory hierarchy, SRAM and external-memory use, interconnects, data paths, power, and thermal requirements.
Optimize the system across heterogeneous hardware, from data-center accelerators to resource-constrained edge devices.
Contribute to open-source projects, technical documentation, and a reproducible benchmark plan.
Requirements
15+ years of total experience, with substantial experience in AI systems and distributed learning.
5+ years of professional Rust development, focused on distributed systems, peer-to-peer networking, or protocol implementation.
Strong understanding of networking protocols and hands-on experience with technologies such as TCP/IP, UDP, WebRTC, or libp2p.
Experience building scalable, fault-tolerant distributed systems.
Practical experience with model sharding and distributed AI workloads; familiarity with federated learning and distributed inference.
Experience with modern AI serving architectures, including prefill/decode disaggregation and the movement or placement of model state such as KV cache.
Strong understanding of security, fault tolerance, consensus algorithms, blockchain integration, and cryptographic techniques.
Direct experience with accelerator or chip-level architecture and optimization. You can engage with hardware implementation and performance trade-offs, not only pass requirements to a hardware team.
Ability to prototype, profile, test, and explain technical decisions with evidence.
Preferred
Experience with WebAssembly (WASM) and cross-platform Rust development.
Familiarity with decentralized AI, swarm intelligence, and peer-to-peer compute networks.
Experience with Docker and orchestration frameworks such as Kubernetes.
Contributions to relevant open-source projects.
Co-founder terms
Compensation is equity only at this stage; no cash salary is currently available. Equity and vesting will be discussed based on the co-founder’s commitment, experience, and responsibilities.
If you’ve built distributed AI systems at the protocol level and can connect networking, model execution, and hardware architecture, we’d like to hear from you.

Work arrangement
Yes