Shipping the World's First AI-Native Dating App while Achieving a Patent in the Process

  • 01
    Project Overview
  • 02
    The Challenge
  • 03
    Our Approach
  • 04
    Results & Outcomes

Project Overview

Picme is the world’s first AI-based social and dating app supported by patented technology and designed for Gen-Z. The founders of Picme are highly talented product and business gurus who needed a technical partner to bring their vision to life. Akava partnered with Picme on all facets of product development: user research, design, architecture and development. We delivered multiple iterations; from initial prototype through a feature robust MVP.


The Challenge

Complex Technology

The overall architecture and supporting backend technology needed to power the Picme platform was very complex and required superior subject matter expertise. We had to create several test scenarios to baseline and predict future outcomes in an effort to implement machine learning tech while being mindful of the vast scalability of the overarching application architecture.

Data Accuracy

One of the biggest value propositions of Picme is the ability to extract metadata found from within pictures of users’ mobile camera roll to generate an accurate portrayal of their likes, interests, hobbies etc. Ensuring a high degree of accuracy (90%+) was paramount as this validates the overall UVP of the platform.

Processing Speed

Automated profile creation is another unique value proposition of the Picme platform. To ensure a pleasant user experience, we needed to ensure we could create a profile behind the scenes while a user was distracted by basic app on-boarding tasks. We had to process and categorically assign thousand points of metadata from up to 15K photos within minutes.



Our Approach

We came up with a thorough plan to deal with the problems at hand.

Alignment & Execution


Given the complexity of the platform, we had to conduct a fair amount of research around the latest/greatest machine learning OSS projects to see what could be leveraged and adapted off the shelf to conserve time and expenses. We developed prototypes and conducted extensive testing of different scenarios to identify potential bottlenecks and challenges in the system. Along the way, we thoroughly documented the architecture and technologies so we could further advance the platform as new edge cases were introduced. We implemented scalable solutions to accommodate potential growth in user base and data volume.

Data Precision


We utilized advanced machine learning models for image analysis and metadata extraction to improve accuracy. In addition, we implemented data quality assurance processes to validate and refine extracted metadata, ensuring a high level of accuracy. Our feedback loops were tight and consistent so as to continuously improve the accuracy of the models based on user interactions and corrections. We introduced several new metadata categories testing them with the machine learning models so we could keep them up-to-date in an effort to perpetually improve.

Parallel Processing


Given speed and accuracy were the name of this technical game, we had to implement parallel processing techniques to distribute the workload to increase overall processing speed. We had to also deploy asynchronous processing for background tasks, allowing the app to perform profile creation while users focus on onboarding tasks. Caching mechanisms were a must as we had to store and quickly retrieve frequently accessed data, reducing the time required for repetitive tasks. Over time, we optimized algorithms for metadata processing to reduce computational complexity and improve efficiency. Lastly, we employed load balancing strategies to evenly distribute processing tasks across available resources thus preventing bottlenecks.

Tools & Tech Selection

The project utilized a range of technologies to achieve its goals, including:


Leveraged and adapted OSS version to meet robust needs.

AWS Cognito

Primary authentication layer with identification federation.

AWS Dynamo

Autoscaling for workload management while being highly performant.


Dynamic scaling while future proofing for container orchestration.


On device, native image processing.


Cross platform mobile UI capabilities.


Results & Outcomes

The outcomes of the project were profound:

App & Platform Delivery

From the onset, the founders of Picme were aware of the technical hurdles that their dream entailed. However, we managed to overcome these challenges by developing new skills and pushing the boundaries of technology. After numerous trials, tests, and constant enhancements, we launched an exclusive MVP to a community of 40K students, which has received considerable acclaim and enthusiasm.

Awarded Patent

Through the development of the advanced technology paired with the product vision of the Picme team, the core algorithm was awarded a patent deemed under a “MOBILE GAME USING IMAGE EXTRACTION”. Further, with Picme team is currently in the “PATENT PENDING” phase for another groundbreaking patent submission. This one is geared towards safeguarding the Intellectual Property for auto-generating profiles on mobile apps, using data directly from mobile devices.

Secured Funding

Picme is pioneering a new era in social dating, leveraging unique and patented technology to redefine the landscape. Our instrumental role in envisioning, designing and developing the product has sparked an outpouring of investor enthusiasm. To date, Picme has successfully raised over $600K – a testament to its potential. What’s even more remarkable is that it’s the investors themselves who have sought out Picme, eager to contribute to this revolutionary venture.

Partnering with Akava was an exceptional journey from concept to reality. Their team transformed our ambitious vision into a first-class platform, brilliantly blending business acumen, engineering prowess, and product-oriented focus. Akava’s tech expertise, creativity and efficiency surpassed our expectations; they are truly technology polyglots and artisans!

Erick B.

CEO & Founder