CDFAM Washington DC, July 15-16, 2026
When Ryan McClelland stood on stage at CDFAM NYC in 2024 and introduced the concept of ‘Text to Spaceship‘ for NASA, it was wildly ambitious given the state of ‘Text to CAD’ at the time (a garbage pile of useless meshes), but at the same time deeply prescient of how the future of design and engineering is shaping up today.
The idea that you could move from a written Systems Requirements Document, to a mission-ready hardware design ready to launch into space through AI-driven computational workflows seemed closer to speculative fiction than a reality.
But, two years later, Ryan is back on stage at CD/DC as keynote speaker, to show how much closer to reality it has become.
This is how the CDFAM program (and industry) has evolved in 2026: From a somewhat disconnected array of ideas, software and data formats, various players have spent years advancing individual capabilities that are now beginning to come together into systems that solve problems, not individual solutions.
This same theme will be explored in an evening session for invited government observers that will showcase how an interconnected workflow between six different companies, and their tech stacks can combine to solve complex engineering problems in near real-time with a coherent data strategy and AI augmentation.
A live demonstration will show how fully interoperable, AI-native tools and agents can integrate new sensor and compute capabilities, update an entire UAS design, and re-certify it for manufacturability and safety in less time than it takes to assemble an Ikea bookshelf.
The broader program reflects where computational design actually sits across industry, academia and government right now.
Brandon DeMille from General Atomics Aeronautical Systems will present on topology optimization for aerostructure conceptual design, specifically how computational methods compress the time between initial concept and informed decision.
Rebeka Melber from Istari Digital addresses the organizational problem: executing the digital thread across a real defense industrial base where every partner uses different tools and operates at a different level of digital maturity.
Steve Massey from SysGit brings the argument that what transformed software development was not (just) better code, but version control as infrastructure, and makes the case that hardware engineering is overdue for the same transition.
Juan Alonso, co-founder and CTO of Luminary, will show what happens when Physics AI surrogate models replace traditional meshing and solving for aerodynamic and structural analysis. Evaluation cycles that previously took days compressed to seconds, across UAVs, submarines, and supersonic missiles.
Pratap Ranade from Arena Physica returns to CDFAM, this time with Atlas RF Studio in public beta, moving AI-driven electromagnetic design from research demonstration into practitioner hands.
Some of the most forward-looking presentations on the program come from directions (and forces) that may be less intuitive, and may not have immediate applications, but are worth considering as we start to move through continuous differential fields and past static material systems.
Giorgia Cannici from Virginia Tech will present on magnetotropic plants: organisms engineered so that their gravity-sensing organelles respond to magnetic fields rather than gravity, allowing growth direction to be actively controlled in real time through externally applied fields.
Matthew Shomper from Not a Robot Engineering returns with similar argument from a different direction: the microarchitecture of bighorn sheep horn as a model for impact protection geometry that no conventional manufacturing process could previously produce, now accessible through additive manufacturing.
Alexander Htet Kyaw from MIT connects generative AI and robotic assembly, addressing the gap between a 3D model and a buildable, physically composed object.
Marta D’Elia from Atomic Machines describes their Matter Design Engine, a workflow that converts informal descriptions into manufacturable MEMS 2.0 devices while using as little AI as necessary, grounded in physics, constraints, and verification throughout.
The architecture and structural engineering sessions reflect an industry that arguably has a much deeper history, and expert base in computational design.
Cooper Schilder from Stantec presents a simulation-first, ML-augmented framework for curved steel beams that separates physics-based analysis, explicit fabrication logic, and learning-assisted ranking into transparent, auditable layers.
Sergey Pigach from CORE studio at Thornton Tomasetti is back at CDFAM presenting on multi-agent AI collaboration and MCP-connected CAD, describing the practical reality of integrating agentic systems into an engineering firm at enterprise scale.
Damola Michael and Elliot Glassman from CannonDesign present Aurora, a pre-design climate intelligence platform that brings EPW data, CMIP6 projections, and ASHRAE conditions for year-round site analysis well before any massing or conceptual renders are pumped out with the latest diffusion sorcery, disconnected from physics or reality.
Brian Ringley from Boston Dynamics approaches humanoid robotics as a question of teaching rather than programming: what does it take to build the AI brain that makes hardware generically capable of interpreting and responding to the physical world as it actually is, not as it was modeled.
This is an adaptive design systems concept, much more than ‘just’ a robotics solution.
Matthew Ellis from Neural Concept, Francis Bitonti from Lexset, Andrew Acuff from SimScale, Alexander Lavin from Pasteur Labs, Rik Baruah and Neel Goldy Kumar from Intact Solutions, and Rhushik Matroja from Cognitive Design Systems round out a program that spans electromagnetics, synthetic training environments, agentic simulation workflows and AI-native physics
A few additional sessions are still being confirmed and will be announced shortly.
You should definitely join us. Space is limited by venue capacity.
Early bird registration is open until June 1.
Government employees can register to attend evening sessions at no cost.
