A Singapore Government Agency Website
How to identify
Official website links end with .gov.sg
Government agencies communicate via .gov.sg websites (e.g. go.gov.sg/open). Trusted websites
Secure websites use HTTPS
Look for a lock () or https:// as an added precaution. Share sensitive information only on official, secure websites.
LogoLogoHomeAboutFAQsEventsProblem Statements
LogoLogo
Sign up here

{build} Hackathon & Incubator

Are you ready to be part of the next {build}?

Contact UsReport VulnerabilityPrivacy StatementTerms of Use
GovTech 10th AnniversaryGovTech 10th Anniversary

© 2026 Government Technology Agency of Singapore | GovTech

Projects/Developer Tooling
GPUmonkeys

GPUmonkeys

Streamlining AI development with our one-line Python integration that provides instant, cloud-agnostic GPU access - letting developers focus on building, not infrastructure management

Booth DT1

GPUaaS 🐒

By GPUMonkeys 🙈🙉🙊

Members 👥

  • Teng Fone 🙈
  • Elizabeth 🙈
  • Wei Xuan 🙉
  • Jian Wei 🙊
  • Willy 🐵
  • Emily 🐒

Problem Statement ❗

AI developers and engineers struggle with platform restrictions, complex setups, and limited GPU access, hindering flexibility and productivity.

Existing solutions lock users into isolated environments, making it difficult to work seamlessly across devices.

Our solution provides cloud agnostic GCC GPU access with a simple one-line Python integration, enabling full control, minimal setup, and unrestricted AI development from any device or environment.

Our Solution 💡

Our team proposes developing a serverless GPU computing platform that enables developers to execute AI workloads with minimal setup.

By integrating a Python SDK, users can define and deploy functions that automatically scale and utilize GPU resources as needed.

This platform will support various AI tasks, including model training, inference, and data processing, without the need for manual infrastructure management.

Comparison 🔄

AspectCurrent PracticeWith GPU-as-a-Service (GPUaaS)
Platform SelectionDevelopers must choose between AWS, Azure, GCP, or on-prem GPUs before starting AI/ML work.No need to commit to a specific platform - GPU resources are dynamically allocated and cloud-agnostic.
Authentication & AccessRequires authentication and setup on each cloud platform separately.Single token-based authentication provides seamless access across platforms.
Setup ComplexityManual configuration of environments, dependencies, and networking for cloud GPUs.Minimal setup—install SDK and run AI/ML workloads with one command.
Hardware FlexibilityLocked into specific cloud providers or need to invest in physical GPUs.Aggregates GPU resources from multiple cloud providers and allows registration of personal/on-prem GPUs.
Deployment & ExecutionDevelopers must manually manage workload distribution, scaling, and cost optimization.Automatically provisions GPU resources based on workload demand, optimizing cost and performance.
Mobility & ScalabilityRunning workloads across multiple devices requires extensive setup and configuration.Seamless execution from any device without platform restrictions, improving mobility and scalability.

Demos 🎮

  • Llama3 8B demo
  • Stable Diffusion Demo
Back to all projects