Building an AI powered product in 10 days
Time was the biggest constraint. I prioritized core mechanics like smart exercises, real time feedback, and clean UX. AI helped generate the content and initial design drafts so I could focus on building and testing.



Creating diverse, intelligent exercises
I built dynamic prompts that allowed GPT to generate a wide range of tasks based on topic, level, and user performance. This included fill in the blanks, listening practice, error correction, and short translations.

Designing with AI collaboration
The entire UI was based on AI generated wireframes and design prompts. I used Midjourney for layout inspiration and Figma to finalize everything. GPT helped organize user flows and microcopy.


Fast and natural localization
DeepL API was used for auto translation of the full interface. GPT then refined the text for clarity and cultural tone. This enabled fast deployment in Brazilian Portuguese with a native feel.

Disconnected teams and workflows
The company’s sales, analytics, and engineering teams worked in silos. We built one connected flow where AI generated structured outputs that moved between departments with zero information loss. 



Slow and manual scoping
Sales teams were spending up to 15 hours per lead preparing documentation. With Blaster, AI captures client needs, builds a draft specification, and hands it off to a human expert for quick validation.

Lack of visibility for clients
Clients could not see what was being built until the final delivery. We added real time previews, AI generated UI flows, and a live demo environment where clients can comment and review progress throughout the process.

A core challenge in Iterius was building a platform that offers deep process flexibility without sacrificing clarity. Each company can fully adapt workflows, analytics, and team structures to their needs, yet everything remains intuitive and transparent. From hiring pipelines to timesheet logic, all modules are easily configurable through a clean admin layer.

Performance was another priority. With AI-driven matching, automated scoring, and real-time dashboards, we optimized data handling with background processing, lazy loading, and clear UI responses, ensuring a smooth experience even at peak loads.

We also placed a strong focus on system architecture. By designing loosely coupled functional blocks, our team ensured that new modules could be added or scaled independently, reducing risks and speeding up delivery across sprints.

Creating Glowify presented unique challenges that pushed our team to innovate. One major challenge was designing a smooth booking experience to make scheduling as easy as possible. We achieved this by implementing a guided booking flow, allowing clients to select times, preferred specialists, and confirm appointments in seconds. We conducted multiple rounds of user testing to ensure simplicity and ease of use.

The second challenge was handling the complex scheduling requirements across multiple locations and service providers. We developed a robust scheduling algorithm that adjusts dynamically based on location, availability, and client preferences. Collaborating closely with the Glowify team, we ensured this system was flexible and scalable to support their growth. As a result, Glowify now has a responsive platform capable of managing a wide network of specialists and clients across various cities.

A significant technical challenge was ensuring the seamless integration of multiple payment options, especially given the mix of traditional bank transfers, credit/debit cards, and cryptocurrency. We structured the backend to support multi-payment processing, leveraging Felix Pay and Stripe APIs to secure transactions across borders. Each transaction type required unique handling logic to ensure reliability and security, and we implemented real-time monitoring to quickly identify and resolve any payment issues.

Another area of complexity was setting up role-based access across the platform. JetBridge Travel has three main user roles (Client, Admin, and Super Admin), each with specific permissions. We implemented a secure, JWT-based authentication system and role-based access controls to ensure data security and user-specific feature access. Each role’s permissions were carefully mapped out, allowing for flexibility and control across mobile and web applications, while also ensuring compliance with global privacy standards.

One of the primary challenges in Jeanize Denimify was the integration of complex body measurement data into a machine-learning model that could accurately predict the best jean fit for each individual. To tackle this, our team collaborated closely with Revelle Nation’s experts to define key measurements and body shape characteristics. We used these insights to train a custom recommendation algorithm that maps body profiles to ideal jean fits. By iterating on model accuracy and testing with real users, we achieved a highly personalized and user-friendly recommendation system.

Another challenge was ensuring a consistent and visually appealing experience across all devices. With a focus on mobile accessibility, we designed the front end with responsiveness in mind, enabling seamless usage on both desktop and mobile. Additionally, the team paid close attention to data privacy, given the personal nature of the information collected. We implemented strict security protocols and anonymized user data to protect privacy without compromising the user experience.

The Event Horizon project faced several challenges that tested our team’s adaptability and expertise. One of the main hurdles was designing an intuitive yet scalable user experience for both the web and mobile platforms. The solution was to create modular components using React and React Native, ensuring smooth transitions between desktop and mobile environments. We also utilized a custom-built API to ensure that both platforms could communicate seamlessly.

Another challenge was ensuring that event ticketing and payments were handled securely, especially with the integration of Stripe for transactions. Our team carefully designed the backend to include robust validation and data encryption, ensuring a secure, smooth purchase process for users.
Moreover, balancing the user experience with event management functionality was key. We worked closely with the UI/UX designer to iterate on the admin panel, ensuring that it was as efficient and user-friendly for event organizers as it was for attendees.
The result was a highly efficient system that maintained stability even as user activity increased, with the backend easily handling multiple concurrent events and transactions.

Redesigning an app like Clique presented several challenges, primarily related to optimizing user experience for real-time, in-person social interactions. Our initial hurdle was updating the app’s interface to be intuitive and accessible for diverse user demographics. We conducted multiple user-testing sessions, refining design elements, navigation flows, and interaction points to enhance ease of use. Additionally, ensuring that the app could handle high user traffic without lags was crucial. Our developers worked on optimizing the backend structure, resulting in faster load times and smoother performance.

Another significant challenge was building an engaging onboarding process that captured Clique’s purpose without overwhelming users. Our solution was a dynamic, step-by-step tutorial that introduces features gradually. This approach kept the onboarding process light yet informative, which, based on feedback, positively impacted user retention rates. By balancing functionality with ease of use, we successfully launched an app that resonated with both new and returning users.

The NeuroEase project had its challenges, particularly when it came to creating a seamless and intuitive user experience for both children and caregivers. One of the first obstacles we faced was ensuring that the app was simple enough for children to interact with, while also offering deep functionality for parents. Our solution was to develop an interface with a visual-first design, ensuring that both kids and adults could navigate effortlessly.

Another challenge was building a real-time monitoring system that allowed parents and caregivers to interact with the children remotely. This required integrating the Goally Care Team feature into the app, allowing live updates and reward systems. We focused heavily on building secure, scalable communication channels, using WebSockets to deliver real-time information without delays.
Additionally, ensuring data privacy and security was paramount. Given the sensitive nature of the app, we implemented strict security protocols, including encryption for user data and secure payment processing via Stripe, providing parents with peace of mind.
Despite these challenges, the team successfully delivered a feature-rich platform that improves everyday routines for neurodiverse families.

Creating LearnHub was a journey with challenges. Our first hurdle was optimizing the platform for Android tablets while ensuring a responsive, intuitive interface. We achieved this by adhering to Material Design principles and rigorous user testing with educators and students.

Another significant challenge was designing the admin console. It required powerful features without compromising user-friendliness. By employing modular design and Python-based APIs, we created a flexible, scalable system that adapts to diverse school requirements. Regular collaboration with stakeholders ensured we delivered exactly what users needed.