For this month’s Five Questions interview, we chatted with ARM Member Matt Brown, the CEO of ThoughtForge! In this piece, Matt explores his background, gives an update on ThoughForge’s ARM Institute funded project, details the value that ThoughtForge has found in ARM Institute membership, and more.

1. We always like to start our interviews by exploring our expert’s background. Can you tell us about your role at ThoughtForge and how your career journey led you there?

I’ve spent over 24 years in AI, driven by a fascination with the nature of thought and how living systems generate agency. Before founding ThoughtForge, I worked on the Aegis Combat System at Lockheed Martin and built AI systems for video game franchises such as Halo, Bioshock, and Splinter Cell. After leaving the game industry, I helped pioneer early applications of Deep Reinforcement Learning at Bonsai AI (now part of Microsoft). ThoughtForge was born from a desire to develop biologically-plausible learning systems—grounded in cutting-edge neuroscience—that empowers robots to interact with the world as living systems do: efficiently, dynamically, adaptively, and resiliently.

2. Can you tell us a bit more about ThoughtForge and why ThoughtForge joined the ARM Institute?

We joined the ARM Institute because the ARM Institute’s commitment to pioneering robotics, automation, and digital manufacturing perfectly complements our mission. The Institute offers a unique collaborative platform where industry leaders and researchers push the boundaries of technology, and we saw a great opportunity for the community to contribute to our human-centric approach. By uniting with the vibrant ARM community, we’re tapping into a wealth of cross-industry expertise to refine and advance our innovative learning systems. This partnership enables us to integrate diverse, real-world insights into our development process, ensuring that every breakthrough not only pushes technological boundaries but also creates sustainable, transformative change across all industries.

3. You have a current project that you’re working on with the ARM Institute called Adaptive Robotic Insertion of Automotive Parts using Multi-modal Artificial Intelligence (AI). Can you give us an update on this project and explain what industry problem you’re solving?

We are addressing the challenge of automating the manual insertion of soft / deformable parts into rigid objects for industries like automotive, aerospace, and electronic manufacturing. Currently, human operators perform complex, repetitive insertion tasks—such as assembling vehicle door latches—with high reliance on manual precision and force adjustment. Conventional robotic solutions struggle with the uncertainties and complex contact dynamics involved in these tasks, making them inefficient and unreliable. Our project uses advanced AI and sensor integration to develop a robust, adaptive robotic insertion system that addresses these issues and can be applied across industries.

We’ve successfully developed our preliminary multi-modal control model, incorporating Force-Torque sensors to guide the insertion process, as well as vision models, and are partnering with Siemens bin-picking solutions for accurate part handling. Currently, we are fine-tuning and testing the models to ensure optimal performance before integrating the components into a fully functional, adaptive robotic insertion system.

4. What advice do you have for other ARM Institute Members who are interested in getting involved with ARM Institute-funded projects? How are ARM Institute projects different from how you would normally operate?

I think our entire leadership team would agree that the ARM Institute and its community members have been some of ThoughtForge’s most active, vocal supporters and strongest accelerators to market. Joining the ARM Institute is one of the best decisions we’ve made as a business to date. As a new technology company, the biggest challenge a business has to acquire customers is not building the technology or even getting POCs, it is getting public 3rd party validation of the value and performance of the technology. Companies who work with start-ups choose to work with them because they see that start-up as offering a unique competitive advantage. Naturally and understandably, those companies want to keep the work and testing they’ve invested in with that start-up private. With an ARM Institute project, you know what you’re building is valuable to a broad set of customers because the customers are the group deciding to fund the technology, and 3rd party validation and promotion of the results of the funded project are built into a 400 company network of your target customers. This is a huge accelerator for any new technology, and especially for a company like us that is doing something completely different from everyone else in the market.

5. What are a few things that you’d like for people to know about ThoughtForge/what exciting things are you working on?

We’re pioneering efficient, real-time multimodal learning and adaptive AI at the edge for dynamic and dexterous tasks like assembly insertion—one of the most critical and unsolved challenges in industrial automation.

While researching for a talk in December, we came across the ARM Institute’s founding deck from 2017. The NDIA/ARM Institute report identified assembly automation as a key challenge, stating:

“Product assembly remains as the last frontier to benefit from process automation with many companies assembling their products in low labor wage countries due to the complexity required and lack of automated systems to grasp and manipulate unique parts. Overcoming these challenges will bring back many manufacturers to assemble their products in the United States.”

This challenge is precisely why I developed the AI that powers ThoughtForge.

Our 2024 ARM Institute project with Siemens and Magna addresses one of the most complex and historically difficult problems in robotic control. Traditional automation struggles with the high variability and precision required in real-world assembly environments—something our adaptive AI is designed to overcome.

We’re excited to share the results of this project with the ARM community, providing a strong validation of ThoughtForge’s AI-driven approach to real-time robotic adaptability.

Join the ARM Institute’s 450+ Member Organization Consortium