Roan Weigert spends his days translating complex AI technology into something developers can actually use. As DevRel AI Engineer at Aparavi, he builds tutorials, hosts workshops, organizes hackathons, and gathers the kind of user feedback that shapes how products evolve. His path to this role cuts through video production, data analytics, and years of building communities around emerging technologies.
The Work at Aparavi
Aparavi tackles a problem that plagues most organizations: unstructured data. Weigert estimates that 80% of company data sits in documents, PDFs, emails, logs, images, and videos scattered across systems with no coherent structure. The information exists, but accessing it remains difficult.
AI can help unlock that data, but there’s a prerequisite. Before any model can deliver accurate results, the data layer needs attention. Aparavi handles discovery, verification, classification, and preparation of unstructured data so AI systems can process it effectively.
Weigert’s job sits at the intersection of that technology and the people who need to use it. He creates example projects and product documentation, runs events, and maintains constant communication with the developer community. The feedback he gathers goes directly to product teams. “My job basically is to make the technology easier to learn, to use, and to adapt, while making the product more useful,” he says.
A Non-Traditional Path
The combination of computer science, video production, and data analytics sounds eclectic, but Weigert sees them as complementary. Computer science gives him the technical foundation to understand and build features end-to-end. Data analytics lets him explain the value of insights. Video production taught him storytelling and how to communicate clearly, whether on camera or speaking to hundreds of people at an event.
He has produced over 1,500 videos throughout his career. That experience shows up in unexpected ways. When reviewing client project requests, he notices they arrive better organized than before. AI tools help clients brief projects more clearly before the first conversation even happens.
AI handles repetitive production tasks well. Weigert uses it to review podcast footage and identify moments likely to perform well on social media, saving hours of manual review. But the creative decisions remain human. “Humans still win the taste, the storytelling, the empathy, and knowing what actually matters to the audience,” he says. “AI helps but doesn’t replace their intention.”
Lessons from the Podcast
Weigert hosts AI Inside San Francisco, a podcast where he interviews AI founders and researchers. After more than 22 conversations with CEOs, data scientists, and directors, he notices patterns in how the field has evolved.
In late 2023 and early 2024, uncertainty dominated. People expressed confusion about AI’s future and struggled to grasp its potential. By 2025, significant progress has been made, though his guests still acknowledge that the field hasn’t figured everything out.
One observation comes up repeatedly: technology advances exponentially while society’s understanding moves linearly. That gap creates friction, and it keeps growing. Weigert sees part of his role as helping close that gap. “It’s one of my jobs and my peers’ job to make sure to keep people informed, build those communities, and share the knowledge so we have a voice in the AI community for the government and for the big tech companies,” he says.
Building Developer Trust
Developer communities don’t form around marketing promises. Weigert has learned that trust requires understanding developers’ frustrations, knowing the problems they’re trying to solve, and speaking their language. Events and community gatherings help, but authenticity matters more than polish.
Developers want solutions to real problems, preferably with proof. The market overflows with tools that claim broad capabilities but deliver little value. To earn trust, Weigert focuses on working examples, open conversations, and honest acknowledgment of failures.
“They need to comprehend that we are on their side,” he says. “We’re not there to try to force a sale. We are there to help them succeed, and if it fits their needs, we offer our product for them to test it out.”
Where Companies Go Wrong
Weigert sees companies make similar mistakes when adopting AI. The rush to add AI features often outpaces any clear understanding of the use case. Organizations want to use whatever data they have, but that data is frequently messy, insecure, or duplicated. The results disappoint.
The problems multiply from there: private information leaks, missing contacts, files scattered without structure. AI performs poorly on bad data.
His advice starts with understanding the specific value the organization wants to achieve. From there, the work involves discovering the right data, establishing proper governance, and structuring what was previously unstructured. Only then can AI deliver meaningful outcomes.
Community as Purpose
Weigert has founded companies in drone technology, video production, and other industries. Those experiences shaped how he thinks about what motivates people. “I understand that money is not the only thing that moves people,” he says. “It’s actually the excitement and that energy to put people together for conversations and discussions.”
He finds building and maintaining communities fulfilling, particularly during what he calls “this new era of uncertainties.” The work offers opportunities to bring others into the field and create space for conversations that might not happen otherwise.
The Future of Developer Relations
AI creates new capabilities, but it also creates confusion. Weigert believes the need for people who can translate, simplify, and connect technology to real outcomes will only grow. Developer relations roles may become what he calls “the human interface between products and people.”
For anyone from a non-traditional background looking to enter the AI field, his advice is direct: start using the tools. Watch videos, read articles, become a user first. Once comfortable, build tools that help others.
“Right now is the best time for you to become a coder without coding skills,” he says.
