JJ Eaton
JJ Eaton is an independent software engineer, maker and AI engineering mentor who builds pragmatic SaaS, micro-SaaS, local-AI systems and custom software. Through BuildThings, he helps businesses apply AI through workflow automation, custom software and private/on-premise AI.
At a Glance
- Growing businesses needing practical AI automation
- Founders and teams needing product-minded software execution
- Businesses with sensitive data that cannot use public-cloud AI by default
- Companies with local IT partners that need software and on-premise AI implementation
- +1 more
AI Tools by JJ Eaton
(1)DeepSeek-V4.1 TensorFold Spark
DeepSeek Serving on DGX Spark
Discussions
No discussions yet
Be the first to start a discussion about JJ Eaton
Latest News
Published results for Compute Lanes, reporting 6.3-second replies versus 26.9 seconds and 78% fewer computation tokens on some tasks, with no gain on standard benchmarks.
Published an explanation of why configured AI agents can outperform a single smarter model by reducing token waste and standardising roles.
Announced an early Compute Lanes feature allowing a local DeepSeek setup to run code while it is still answering; the post reports 38 of 40 coding tests solved and a 16% overall speed improvement.
Compared one 512GB Mac Studio with two 256GB machines and argued that two smaller local-AI systems can provide more compute, parallelism and resilience for similar spending.
Products & Services
Perth-based AI consulting and software business offering workflow automation, custom AI software, on-premise/local AI and hardware, AI-readiness workshops, ongoing support and a free website-security assessment.
An open-source recipe for serving DeepSeek-V4.1-Flash on two NVIDIA DGX Spark systems with the TensorFold engine. It includes an EXL3 pack, native CUDA/Zig serving, benchmarks, configuration and an OpenAI-compatible API path.
A local AI router, scheduler and voice loop for one GPU box, exposed through an MCP server and a lightweight web voice UI. The public repository describes support for local chat, speech, image and video models, loaded on demand.
A browser-based PDF editor for adding text, filling forms, drawing or placing signatures, reordering pages and downloading files locally without signup, subscription, watermark or server upload.
Market Position
Eaton positions his work as a pragmatic, product-minded alternative to large consultancies and generic cloud-AI deployments: direct delivery by a Perth-based engineer, maintainable SaaS architecture, and local/on-premise AI where data custody matters. His open-source TensorFold and LocalRouter work also positions him among independent local-inference and self-hosted-AI builders, competing conceptually with vLLM-based local serving, hosted AI APIs and larger AI-automation consultancies.
Leadership
Founders
JJ (Jayj) Eaton
Software engineer and founder based in Perth. His published résumé records management/consulting work at SpeedEFX (2017–2022), a Software Engineering Immersive at General Assembly (2022), software engineering at Pricemates, Upwork software engineering from August 2022, and software engineering at VGW/Global Poker from April 2022 to July 2023. His portfolio says he has 15 years in business, five years shipping software and 30+ projects completed.
Founding Story
Eaton describes a transition from business management and consulting into software engineering, with an emphasis on combining business context and maintainable implementation. His stated focus is early-stage SaaS architecture and practical AI systems that move quickly without becoming fragile.
Business Model
Revenue Model
Freelance software engineering and fixed-scope/custom AI consulting and implementation through BuildThings, with optional ongoing support. The free SimplifiedPDF and SimplifiedQR tools accept optional Buy Me a Coffee support; BuildThings says audit/workshop scope and pricing are agreed with each customer and private-AI work is sold as a system of software, hardware and model rather than a mandatory subscription.
Target Markets
- Growing businesses needing practical AI automation
- Founders and teams needing product-minded software execution
- Businesses with sensitive data that cannot use public-cloud AI by default
- Companies with local IT partners that need software and on-premise AI implementation
- Developers and enthusiasts operating local AI models on NVIDIA DGX Spark hardware
- Early-stage SaaS and micro-SaaS product development
- AI workflow automation for repetitive business administration
- Custom internal tools and customer-facing AI applications
- Private AI for sensitive legal, medical, accounting and commercial records
- Local inference and multimodal assistants on owned GPU hardware
- Browser-based PDF form filling and signing