Holokai
Before you read...
Due to multiple NDAs associated with this role, I am limited to sharing high level information only. Specific details and visual assets cannot be shown publicly.
Ground Zero
In April 2025, while my contract work with the USFS was being disrupted during the early stages of DOGE, I was approached by the CEO of Dynamo with an unexpected opportunity. He asked whether I had the bandwidth to help with a passion project he was exploring at the time, known internally as Project Nova. The idea was still in its infancy, but the vision was clear: to build an AI driven product that could accelerate the early stages of development when working on contracts.
I accepted the opportunity without hesitation. AI was rapidly becoming part of everyday life, fueled by the rise of tools like ChatGPT and similar platforms, and I was eager to be involved at such an early stage. My role in the project was challenging at first, as it required a significant mindset shift. I had to step away from the structure and caution of government contract work and adopt a startup mentality, one focused on speed, experimentation, and producing tangible results rather than perfecting every detail upfront.
Trials and Tribulations
The team was intentionally small, made up of myself as the Lead Designer, two Project Managers, and four developers. Having two Project Managers was unusual, but the CEO believed it was necessary given how ambiguous the initial requirements were. They worked closely with him to shape and clarify those ideas before bringing them to me in a more digestible form. Eager to make an impact on the project, I reviewed the translated requirements and gave them the green light. In hindsight, this was a moment where I should have slowed down and spent more time questioning and refining the scope before moving forward.
Following the CEO’s direction to focus on momentum, I moved quickly into Figma and began designing. With no established visual direction for the product, I took ownership of defining a baseline look and feel. I created a lightweight, temporary brand system that could evolve over time, including a Dynamo aligned color palette, Dynamo aligned typography, and a design system built on Google Material UI. With this foundation in place, my goal was to ship a V0.0.1 product that met the core requirements and supported a functional CRUD experience.
Using Material UI allowed me to move quickly through the initial mockups, as the primary challenge became finding the right layout rather than building components from scratch. The application was organized into sections that represented each subcategory within a government contracting company, with each section supported by its own AI assistant tailored to that team’s specific tasks. To guide the
overall structure, we drew inspiration from the layout popularized by Claude, adapting key elements to fit our use case while maintaining a familiar and intuitive experience. The layout I landed on felt like a strong starting point. It was modern, aligned with the temporary brand, and captured the core requirements while remaining flexible enough to evolve without major structural changes. At the time, it felt like the right solution. Looking back, however, this was the moment where I made another important mistake.
Rather than stopping there, I continued designing additional screens to illustrate how users might move through workflows within the application. While some requirements guided these flows, many decisions were based on assumptions I made at the time, with the expectation that they would be refined and adjusted after reviewing them with the team.
After building out four separate workflows, I decided it was time for a review. Two weeks of focused design work had me eager to share my progress with the team. The Project Managers and I went through the designs together, making minor tweaks to ensure everything aligned with the requirements. Overall, they agreed the workflows were ready to present to the full team.
A week later, we scheduled a session with the entire team, including the CEO. I spent about thirty minutes walking everyone through each workflow, explaining functionality and pausing for questions. What felt like an eternity of silence ended when the CEO unmuted his mic. He began by praising my work, which was a relief, but quickly shifted to asking the developers about their progress. Since this was the first time the developers had seen the designs, they had not started any UI work. Frustration came through strongly as the CEO emphasized the need to produce tangible work rather than spend too much time on pre-work. He closed by reiterating that he liked my designs but wanted to see more output.
I left the call feeling defeated, carrying the weight of having let the team down. It took about a week to gain clarity from the Project Managers on the path forward. The result was a major shakeup: a new executive was brought in to lead the project, additional developers joined from Vietnam, several existing developers left, and one of the Project Managers stepped away from the team.
New Beginnings
About a month had passed since I last heard about Project Nova. I was wrapping up work for the Permitting Council, continuing projects with USGS, and returning government equipment from my Forest Service assignments. It was a period of uncertainty, moving from a workload that had been full and engaging to just a few projects in their final stages. In the midst of this, I was contacted by one of the executives at Dynamo.
They were looking for help building a new brand. Dynamo was launching a child company, Holokai, to serve as the face of all their AI initiatives. At the time, the company had no identity at all. The executive shared a collection of screenshots they liked, which would serve as inspiration for the visual direction and overall vibe that Holokai should embody. In the back of my mind I was still wondering what was going to happen to the product that was once named Project Nova, but I was just happy to continue working on this project so I moved forward.
They clarified that I didn’t need to work on a logo, as graphic designers were already handling that. My primary focus would be building the brand while simultaneously designing a website. The team assembled for this project included me as the lead designer, a Project Manager, and a Developer the same team I had worked with on Permitting Council, a government sector brand built from scratch. I enjoyed collaborating with this group and already felt comfortable leaning on them for guidance and feedback.
In our first meeting, the developer suggested trying a new design system called Tailwind UI. I had never worked with it before, but I welcomed the opportunity to step outside my comfort zone with Material UI and USWDS. After some research and discussions with the executives, we decided to move forward with Tailwind UI for its wide range of preset components, modern aesthetic, and seamless Figma integration.
With the system in place, I began developing branding options. Each explored different typefaces and color palettes, and I included a stock Tailwind component landing page alongside these options to help visualize how the design could look and feel in a real website context.
I met with the executives to review the initial branding options, and we went through several rounds of iteration. Choosing the primary color was straightforward, but there was more discussion around the secondary and accent colors as we tried to strike the right balance.
Selecting a typeface was easier. Poppins, the default font in Tailwind, resonated with the executives and quickly became the cornerstone of the brand. After a few days of fine-tuning, we had established a cohesive visual identity that could guide the project moving forward.
While we refined the branding, the Project Managers were gathering requirements for the website. They organized the requirements into two phases: features for the initial launch and those planned for post-
launch, providing a clear roadmap for development. Their primary goal was to establish a web presence as quickly as possible, and the requirements reflected that urgency. Every element of the website was designed to inform users or potential customers about who Holokai is and what the company does.
Using the requirements provided, I was able to create a low-fidelity mockup of the landing page in just two days. Learning from earlier missteps on the project, I checked in with the Project Managers and executives to ensure the design aligned with their vision. The mockup passed review without any changes, giving me a significant confidence boost after the challenges of previous reviews. I then moved on to low-fidelity mockups for the other parent pages including Product Overview, Careers, and About Us, and these designs were approved as well.
When it was time to develop high-fidelity mockups, I ran into some bumps due to the content not being ready. The executives were responsible for providing copy, but they had not had the time to finalize it. To keep momentum, I used the available content and leveraged Figma’s AI to fill in the rest, creating realistic simulations that could still be reviewed.
I held another review with the team, focusing the discussion on layout and overall components rather than the specific content. The executives were eager to move into development and gave the green light. From there, I collaborated with the developer to break each page into Tailwind components, providing a solid foundation before adding custom styling.
I took the requirements and created a low-fidelity mockup of the landing page within two days. Learning from earlier mistakes on the project, I checked in with the Project Managers and executives to ensure the design matched their vision. The mockup passed review without any changes, which was a huge confidence boost after my previous review experiences. I then moved on to create low-fidelity mockups for the other parent pages including Product Overview, Careers, and About Us. These designs were also approved after review with the same group.
When it was time to move the mockups to high-fidelity, I encountered some challenges because the website content was not yet ready. The executives were responsible for providing the content, but they had not had time to finalize it. To keep momentum, I used the content that was available and leveraged Figma’s AI to fill in the rest, creating realistic simulations that could be reviewed.
I held another review with the team to gain executive approval to start developing the website layout. I focused the discussion entirely on layout and general components rather than content. The executives were eager to move into development and gave the green light. I then collaborated with the developer to break each page into Tailwind components, providing a foundation for the layout before adding custom styling.
Launch Week
For about a month, the developer and I worked together to build the website, making sure it matched the intended look, feel, and functionality from the mockups. We still did not have the final content from the executives, so we relied on the placeholder content generated by Figma’s AI. The executives had told us the real content would resemble the placeholders, so we moved forward with confidence. In hindsight, that was a mistake.
We completed the website with two months to spare before the launch date, giving the executives plenty of time to finalize the content. However, a month and a half passed and the site remained unchanged, which was concerning. When we checked in, we discovered the executives had not even started preparing the content. Our check-in seemed to trigger a sense of urgency, and they promised the content would be ready in two days.
When we finally received it, the content arrived as a word document. After reviewing it, we quickly realized that the copy would not fit within the sections we had designed, creating a significant challenge for the layout and overall user experience.
For example, some of the cards had headers that were four lines long. This made the cards excessively tall and gave the site an unprofessional appearance, especially for a startup hoping to attract potential business partners.
We adjusted the content to fit within the design limits and returned it to the executives, only to be met with frustration that the content we had was not acceptable. Our components had been built with character limits in mind, but rather than reworking the copy, they asked us to remove the sections that could not be launched.
This was difficult to accept. The developer and I had worked tirelessly to create a polished, fully functional site in a very short timeframe, giving the executives more than enough time to provide content. With only a few days remaining until launch, we conceded and removed the sections with incomplete content.
The result was a site that did not fully communicate what Holokai was or what it offered. If I were a first-time visitor, I would leave still unclear about the company. While the team at Dynamo was excited to see the site live, I felt both pride in launching the project and disappointment that the final product did not reflect the full potential of our work, even though much of it was beyond my control.
New Faces, Big Dreams
We were shifting from an oceanic, beach-inspired vibe to a wavy, cyberpunk aesthetic, which made sense for a company focused on AI products. It was a visually striking theme that I was excited to explore, and I was equally thrilled
to finally work with an executive who was fully invested in building Holokai. The president wanted to keep the foundational elements of the brand, such as the color palette and typeface, intact, focusing instead on updating the imagery across the website. While we worked on the website revamp, he had a team of graphic designers creating visuals that matched this new cyberpunk-inspired direction.
The new president wanted to completely revamp Holokai’s website and overall brand. Since I had created the original branding, he turned to me to develop new branding options. He shared a set of screenshots that reflected the aesthetic he envisioned for an AI company, and I got to work translating that vision into concepts. Here is what I was given:
Dynamo appointed a new executive as the president of Holokai. Previously, the executives involved had been tied up with other projects, which meant Holokai was not their top priority. The new president’s sole focus was running Holokai, and he approached the role with full dedication. About a month after the website launched, he reviewed the site and, as we had anticipated, was not satisfied with how it presented the company. He was unclear about the purpose of the website and recognized that significant improvements were needed.
Revamp
When meeting with the president, the developer and I shared our concerns about the challenges from the last launch, and he was fully understanding. He planned to hire a dedicated writer to manage all content for the new website, which was a huge relief. His main goal was to make the website more informative for potential customers, something the previous site had lacked. To achieve this, we needed to restructure the site so it was more text-driven while remaining digestible and engaging for users.
The website’s hierarchy also needed a major overhaul. In its current state, the site presented Holokai ambiguously. The president wanted to clearly highlight how Holokai could be used across multiple business types. We researched other B2B AI companies to understand how they structured their websites and used their formats as a baseline. Since we did not have the same volume of content, we selected only the sections that matched the content we could provide, which became the foundation of our site structure.
To manage this content effectively, the developer and I designed reusable sections across the website, which we called Large Content sections. These featured alternating layouts of text and images that could be applied to any page while remaining visually engaging. I created high-fidelity mockups to demonstrate how these sections could be reused without becoming monotonous, keeping the user experience dynamic and consistent.
This example demonstrates how the Large Content sections could be repurposed to showcase how Holokai provides solutions for a variety of businesses and government sectors.
We held several design review meetings with the president, who was enthusiastic about the layout and the information-focused approach of the mockups. He wanted the site to feel more like a polished enterprise website, so to achieve that, we incorporated components commonly seen on these types of sites. One example was a "Who We Work With" carousel to highlight the organizations and sectors where Holokai’s AI solutions are applied.
While creating the mockups, I also drew on my graphic design skills to craft the images that would accompany the text in the Large Content sections. Using Adobe Illustrator, I combined screenshots from the alpha product with graphical elements to achieve a cohesive, tech-savvy look across the site. These visuals were well received by the president, who wanted them applied to every Large Content section. With over 25 sections and fewer than 15 product screenshots, I had to be creative and deliberate in selecting which visuals to use for each section. Through careful planning and design, I made it work, and all images were approved for use on the website.
Over the next two months, I worked closely with the president to ensure every element of the website was polished and ready for launch. This included creating multiple graphics to complement the new content being produced by the writers. Unlike the previous launch, the writers followed strict guidelines for character count, ensuring that each component fit perfectly without appearing broken.
Once the president was confident that the website was complete, we launched. The current site can be viewed at https://www.holokai.ai.
In Summary
In 2025, I navigated two high-stakes AI projects at Dynamo, each presenting unique challenges and opportunities for growth. With Project Nova, I learned firsthand how to pivot from the structured environment of government contracts to a startup mindset focused on speed, iteration, and producing tangible outcomes despite ambiguous requirements. This project highlighted the importance of balancing rapid design with thoughtful review and revealed the consequences of misaligned expectations between design and development.
Later, with Holokai, I took the lessons from Nova and applied them to building a brand and website from the ground up. Collaborating closely with executives and a small team, I established the visual identity, developed layouts, and created reusable content sections to convey the company’s AI offerings clearly. Challenges with incomplete content during the first launch reinforced the necessity of aligning content strategy with design, while the site revamp under a dedicated president demonstrated the power of iterative design, cross-functional collaboration, and creative problem solving in delivering a polished, user-focused experience.
These experiences culminated in several key learnings:
Early alignment on requirements and scope is critical to prevent miscommunication and wasted effort.
Adopting a startup mindset enables faster iteration and tangible progress, especially in ambiguous projects.
Establishing a flexible brand and visual system early allows rapid design while maintaining cohesion.
Placeholder content can sustain momentum, but final content must be incorporated early to ensure design integrity.
Reusable design patterns improve efficiency, consistency, and scalability across diverse content needs.
Close collaboration with leadership and iterative feedback ensures design solutions meet both user and business goals.