JuliaHub provides AI-driven simulation and design solutions for hardware engineering, helping organisations eliminate the bottleneck between initial design concepts and validated, real-world performance.

Its flagship platform, Dyad, acts as an autonomous engineering partner. Paired with leading LLM technology, Dyad can generate, simulate and validate complete physical designs from scratch, integrating rigorous physics-based modelling with AI workflows to automate system simulation while strictly enforcing physical laws.

Dyad is built on the Julia platform and language, now the industry standard for high-performance scientific computing, trusted by over 1,000,000 users across 10,000 companies worldwide.

Julia's origins trace back to 2009, when Shah and his collaborators realised that compilers, not libraries, were the key to solving problems at scale. Libraries, he explains, are trapped in the domain they were built for, while a compiler can operate generally across many domains at once.

That insight shaped JuliaHub's entire trajectory: Julia became a programming language for scientists, and Dyad became a physics compiler for engineers, neither locked into a single niche, both capable of tackling complex, multi-physics, multi-scale problems traditional software could never touch.

Shah describes his early relationship with Julia as intimate, centred on solving the notorious "two-language problem" and proving a technical thesis. As CEO, his focus has shifted to empowering product development teams globally with the technology built by some of the field's leading engineers at MIT and beyond.

He notes that autonomous AI agents now let him stay anchored to the Julia community and codebase even as his role has scaled, a pattern he sees mirrored across many senior technology executives.

Traditional physical systems design is slow and fragmented, with engineers spending weeks manually building mathematical models and tuning parameters by hand. Dyad 3.0 changes this by embedding autonomous engineering agents directly into a physics-based simulation engine.

Engineers can hand the agent plain-language requirements or a PDF specification, and it automatically researches the underlying mathematics and builds the multi-physics model, self-correcting its code to enforce conservation laws and unit consistency and prevent hallucinations.

Powered by Julia and SciML, Dyad 3.0 accelerates simulations up to 100x and can instantly compile designs into embedded C code for physical hardware, delivering up to a 10x productivity boost for industrial R&D.

During a six-week trial with water engineering services provider Binnies, Dyad detected specific pump fault conditions, including early signs of bearing degradation, with up to 95% accuracy.

Rather than relying on separate models for design and operations, Dyad lets a single physics-based model run across a system's entire lifecycle, from engineering and optimisation through to real-time monitoring and predictive maintenance.

For a water utility, catching a failing bearing before it causes a breakdown means maintenance can be planned proactively, avoiding costly emergency repairs and service disruptions, a shift Shah describes as central to building more resilient infrastructure.

Key Takeaways

• →Viral Shah's career-long focus on solving problems at massive scale began with India's billion-user Aadhaar identity programme.
• →Julia was created in 2009 as a general-purpose compiler-based language, deliberately built to avoid the narrow limits of single-domain libraries.
• →Julia is now trusted by over 1,000,000 users across 10,000 companies for high-performance scientific computing.
• →Dyad, JuliaHub's flagship platform, pairs physics-based modelling with agentic AI to generate, simulate and validate hardware designs.
• →Dyad 3.0 introduces agentic simulation, accelerating simulations up to 100x and delivering a 10x productivity boost for industrial R&D.
• →In a six-week trial with Binnies, Dyad detected pump faults with up to 95% accuracy, showing how one physics-based model can serve a system's entire lifecycle.