Industrial Engineering #4 - Carlos Vanegas, Reshape Automation
Guest: Carlos Vanegas | CTO, Reshape Automation
One of the biggest focus areas for Schematic Ventures continues to be robotics and automation solutions. We organize the annual Big Bash event at Modex and Promat, and hosted over 500 warehouse roboticists this year. We’ve put together a robotics hackathon in San Francisco and recently hosted a half-day robotics summit, featuring multiple panels on robotics development and go-to-market strategies, and a startup showcase. Given our excitement for this category, we are so thrilled to be investors in Reshape Automation. We first connected with the incredible co-founders, Juan Aparicio Ojea and Carlos Vanegas, over a year ago when Schematic led Reshape’s preseed round. Since then, they’ve successfully raised a seed round, led by Ironspring Ventures, with additional investment from Bee Partners, Haystack, Supply Chain Ventures, REMUS Capital, and Expansion Venture Capital. I spoke with Carlos Vanegas, CTO of Reshape, to dive further into the tools powering his workflow today, his advice to other startup engineering leaders, and the robotics trends he’s most excited about.
Carlos Vanegas co-founded Reshape Automation to accelerate the transition to automated labor in the manufacturing and supply chain industries across North America. Reshape is building an AI-powered marketplace and platform to help manufacturing and supply chain companies discover, design, deploy, and support the automation that is right for them.
Vanegas also co-founded Synthicity, a bootstrapped software startup in the AEC space acquired by Autodesk in 2015, that built pioneering 3D generative design, simulation, and analytics software applications for urban planning.
He led software teams at Bright Machines during its early years, which built core technology behind modular lines of robotic cells that automate the final assembly of electronic products. He also led data architecture for the Autodesk cloud platform, where he advocated for cloud-centric common data environments that enabled remote collaboration between users and interoperability across Autodesk's flagship applications, and helped shape the data strategy for the company's platform transition.
Vanegas takes pride in building high-performing teams of committed and talented individuals and fostering a culture of trust, accountability, autonomy, and growth. He has led the deployment of cutting-edge software products at both startups and big tech companies across manufacturing, robotics, urban planning, and architecture.
During his PhD years, he was part of several academic research projects in 3D computer graphics and computer vision, focused on semi-automated methods for designing, visualizing, simulating and analyzing cities, and on computer vision methods for 3D reconstruction from photographs and LiDAR data. His research work has been published in top academic journals and conferences, including SIGGRAPH, Eurographics, and CVPR.
Given your extensive background in robotics, what initially drew you to the field?
I’ve always enjoyed spending time working on impactful problems and things that I think will solve either a serious social or economic challenge. Before robotics, I worked on urban planning and design as I loved building technology that makes it easier to design spaces where people live and work.
About five years ago, I became more aware of and involved in conversations about manufacturing and the way we produce our goods. I started learning more about labor shortages and demographic trends that will negatively impact workforce availability in manufacturing, and how these shortages can become one of the most significant challenges we'll face as a nation. I became interested in how we could manufacture and produce our goods in a way that was both environmentally and economically sustainable. How do we empower domestic manufacturing so we can become a more autonomous country and reduce our reliance on foreign-made goods? We don't necessarily want to depend on other countries for our manufacturing capacity so I became more interested in the role that robots should play and how to overcome the bottlenecks that exist today in their adoption, whether it’s technical, commercial, economic, or social.
I had originally liked robots the way I think most people find robots interesting. But over time, I was questioning how we could put robotics at the forefront of manufacturing. How can we enable robots to perform those tasks and jobs that people don't necessarily want to do because they are repetitive or are not ergonomic or create long-term health issues? This returned me to my original purpose of understanding how technology can help make the quality of life better for people and give us more economic autonomy as a country. This is how I initially got into robotics.
I first worked at Bright Machines and they were focused on robotics for electronics assembly and software-defined, self-contained microfactories for manufacturing electronics. This was a really cool idea and I learned a lot from working there.
After working at Bright Machines and then at Rapid Robotics for about five years total, I saw that across the industry, the rate of adoption of robotics and industrial automation among the manufacturing and supply chain industries was quite low, and we needed to speed it up. This was the hypothesis that led to the founding of Reshape Automation and why I joined forces with Juan.
We have theories about the technology, products, and workflows that can accelerate and catalyze the adoption of robots in manufacturing. There has been a lot of progress and advancement but it's still an unsolved problem at the moment.
In your role as CTO at Reshape, which tools and systems have you found to be most helpful in building the product?
As we work to validate product-market fit, we’re running a few experiments at the moment. We’re currently building the Reshape Market, the Reshape Planner, and a new Insights product. These are all ideas we're testing in parallel with customers and design partners and then figuring out which of those are the ones that have the highest probability of making a difference in the way that the industrial automation market operates. In Julian's [Schematic GP] words: “How do we increase the liquidity of a highly illiquid market?” I look at tools and systems in two buckets: those we use in software development, and those we use in our business and operations.
On the engineering front, I'm surprised in general by how much easier it is to build and deploy software these days. A full stack application 10 years ago versus 5 years ago versus today would have significantly different timelines. It feels like every day the infrastructure that we have for software development is getting better. You can build better and faster with a smaller team. Everything from storage to data schemas to CI/CD pipelines to AI agents to frontend libraries has been commoditized. At Reshape we’re using frameworks that, combined with our very talented software team, multiply our development velocity by 10x. To give a few examples, we’re using Remix as our web framework, Flowbite for UI reusable components, Cloudinary for asset management, and Fly.io to host our applications.
We also use Prisma as our ORM, which makes a big difference. I’m a big believer in understanding the systems that we’re trying to model with any software we build, and thinking in terms of Systems Architecture, including the system entities and the relations that exist between them. Describing the system form accurately leads to getting the data models right, which I think paves the way for a much smoother development of the entire app or platform you’re building. Prisma gives us a declarative, version-controlled, and type safe data modeling framework with easy migrations, and it has helped us build an ecosystem of apps, all of them operating on and interacting with common data schemas and entities from day one – a “data at the center” approach from inception. Counting my years in academia, I've been building software for almost 20 years and it's never been as fast as it is now.
On the business and operations front, I’d say there’s a lot of knowledge and information to manage as we speak with customers and run interviews. We’re also managing and creating requirements for our engineering team, and running the go-to-market team to speak with customers and sell the projects and software we're building.
You’ll likely hear a similar answer from many people these days but I'm finding agents, co-pilots, and AI tools to be really useful. For example, we started using Otter.AI to create summaries and tasks from all our conversations. Sometimes I think our minds are not great at keeping track of all these different conversations and ideas and questions that come out of hours and hours of conversations we have every week. Having this copilot to make sure things don't fall through the cracks and highlight our best ideas has been super helpful. We have an integration between Otter.AI and Slack so after every call we get a summary of the call in Slack. This is searchable and it builds an index so I can ask things like, “Have we talked about doing a holiday campaign? If so, what was the decision?”. This could have been from a meeting over a week ago and I’ll get an answer immediately. These tools – like an assistant with superpowers – were impossible just a few years ago but are now much more commonplace. This has been great for more organization and coordination for the day-to-day operations of the team.
Also, I'm obsessed with databases so we use Notion heavily. Everything in Reshape is basically a database. We have databases for everything including customers, milestones, features, people, sprint planning, and our operational tasks. We don’t really use Jira or other dedicated tools; we just have a Notion database and everything is connected so that puts all the company information we manage in one place.
How do you typically find and evaluate new tools before incorporating them into your workflow?
On the software side, we have built a team of great software engineers. We have a few team members who are very well-informed about all the latest frameworks, and the adoption rates and pros and cons of each. When we started building our tech stack about a year ago, and as we continue to add components to that stack, our developers bring recommendations and we review them together. There's no top-down mandate so it's always an open dialogue between the senior developers. So far, we’re really happy with the choices we made, as I mentioned before.
Security is another important factor since we’re working with customers who understandably care a lot about keeping their data protected. As an example, a few months ago we were evaluating different options for industrial VPN routers. We need this so that we can remotely connect to devices that run inside automation projects deployed by Reshape or our partners. Remote monitoring and support are part of the support packages our customers buy. They want peace of mind that the systems they purchase will operate smoothly and that someone is there, physically or virtually, to troubleshoot when things stop working well. We remotely monitor everything that's going on in these automation systems, including the PLCs, industrial computers, cameras, and pretty much any device. We were researching the best solutions for remote monitoring and security was one of the metrics we paid a lot of attention to when choosing a product and vendor for those components in the stack.
Pricing is also a consideration, but usually more so when things start to scale. When experimenting and validating hypotheses in the early stages, pricing is a factor but not a huge one.
Any advice for other startup CTOs building their tech stack from scratch?
I think the non-technical advice here is even more helpful than the technical advice. We’re all engineers and love building but ultimately, what matters more than the technology and the engineering is the product we build. What are the pain points and the real problems we’re solving for our customers?
At Reshape, we embed the team in the user problems so we never build technology in a vacuum. I spend a lot of time at customer sites, which is normal because I'm a co-founder and CTO, but we also bring our engineers to customer sites or conversations. This helps ensure the features we're building and what we're putting into our applications create value directly for our customers and solve real problems and pain points they have. I think that's the only way to build products. Especially at a startup, you have to embed yourself in the customer’s pain points and problems.
Also, you need to set high-frequency milestones. At startups, it’s all about fast iterations, running a lot of experiments, and moving quickly but you should be able to define progress in chunks of value-generating product features or components and then measure against those milestones. This could look like building a feature and then meeting with customers to show them what’s new and get feedback right away. This is more product advice than technology advice but I think as technical founders who are really passionate about the technology itself and its capabilities, we sometimes don’t pay as much attention to understanding what problem the technology is supposed to be solving, and continuously tracking if the technology is indeed delivering on the promise of solving that problem as we build it.
Which trends in the robotics space are you most excited about?
It’s fascinating that science fiction is becoming more like reality every day. Obviously, humanoid robots are cool and they're fun to watch. I was at IMTS a few months ago and it was great to see all the progress companies are making with humanoid robots. I think we may still be far from where the applications can go mainstream and are deployed at scale in real problems. But I'm following this closely because it would be cool to start seeing humanoids in day-to-day situations the same way that all of a sudden it became normal to see self-driving cars in the streets of San Francisco and Austin all the time. Just 10 years ago, this was all science fiction. I think this transformation is going to happen soon and it's going to be exciting.
Additionally, I'm looking forward to seeing more adoption of GenAI in the industrial automation market and workflows. We're seeing GenAI being used for everything from programming robots to simulation tools to creating synthetic data sets for training different types of algorithms.
I also think there's a pretty significant opportunity to apply GenAI and agents to the “less robotic” aspects of industrial robots.
Let’s look at the processes, the matchmaking, the operations, and the transactions: how transactions happen in the industrial automation marketplace, how requirements are gathered, how RFQs are put together, how proposals are created and approved, how projects are planned, tracked and executed, how to evaluate the actual ROI of a half-a-million dollars automation system vs. the promised ROI. Being able to handle these processes and answer these questions accurately, quickly, and continuously is essential to improve the trust, transparency, and liquidity in the industrial automation ecosystem – and this is where Reshape is trying to make a significant difference and where we’re starting to shine.
The platform can support project execution, the operational aspects of deploying products, matchmaking between supply and demand, and adding transparency and liquidity to this market. I think that's where a lot of the impact will be and is something we're working on. This recent article from a16z talked about the role of GenAI and how transformational it will be if we apply it correctly to specific verticals and industries. I don't think GenAI is hype; I think there's a lot of value. We're just starting to explore what the value is going to be.
Any final words or key takeaways for readers?
The transition to automated labor for manufacturers and supply chain companies is an inevitable journey. Labor shortages will get worse and reshoring/nearshoring trends will continue to gain momentum. Just as it has happened with past technology waves, companies that position themselves ahead of the automation adoption curve will emerge as clear winners.
We’ve created Reshape because companies need a one-stop-shop platform for their end-to-end automation journey. We help our clients win by finding and deploying industrial automation with clear ROI and accelerating their transition to automated labor.
So where do you start? Head over to the Reshape Marketplace and start exploring what’s possible. With help from our AI agents, you can browse hundreds of automation solutions and products and discover the ones that are a good fit for your needs. Our team will then work with you to help you get those solutions to your floor and, more importantly, design and execute your automation roadmap.
If you’re similarly excited about building AI-powered solutions for robotics, come join us! Email careers@reshapeautomation.com with your resume and a short intro.
This series focuses on navigating technical software decisions within industrial companies. From optimizing infrastructure choices and leveraging DevOps best practices to harnessing the power of cloud technologies and improving data workflows, our guests will highlight how they've considered these decisions and implemented new solutions across their organization. If you're similarly excited about leveraging technology to empower our national industrial base and/or building solutions focused on this category, please reach out. My email is ananya@schematicventures.com – I’d love to connect!