Enterprise B2B Jewelry Network

The platform’s digital catalogue contains over 100,000+ products and the search process can take a lot of time. To make it the most convenient and simple for the buyers, Leobit’s experts used Google Vision API Product Search and performed integration with Google Machine Learning Service. It allows buyers to simply upload a photo of a similar product they are looking for, and the platform will generate relevant results. The biggest challenge was to keep synchronized huge data sets between Google Cloud and AWS platform required to quickly receive a perfectly matched recommendation on our jewelry enterprise software solution. To overcome this issue Leobit’s developers created an adapter that converts Google responses and sends SQL queries to the system database and keep synced both blob storages. They also added a recommendation cashing layer to optimize waiting time and reduce load from Machine Learning microservice.

We also had to figure out how to improve the performance of real-time price calculations. The final cost of a product is based on a variety of factors such as materials they are made from, materials’ stock prices, currency, discounts, etc. That’s why Leobit experts had to guarantee automatic price updates based on the latest data from suppliers and the stock market. They used Apache Ignite, a distributed in-memory cache, for efficient processing of cached data and accelerated real-time price calculations at 8 times.

Another challenge was to build a self-made real-time scalable and flexible activity tracking. After analyzing the market of existing systems, it was clear that none is fully suitable for all requirements. Leobit’s experts created a self-made solution, which enables to receive information about user actions in the system in real-time. It was built to be generic, actions agnostic, and flexible that allowed to integrate it into various system services, both backend and frontend. This helped to quickly expand the collection of necessary analytics, even live servers.

The legacy part of the system was written in PHP. To ensure high availability, scalability, and easy maintenance, the new version of the system was built on top of the .NET Core platform. As the old version, the application is still supported, and it is getting new functional changes. Data consistency for 2 systems was quite acute and was successfully resolved thanks to the developed data synchronization engine. This made it possible to obtain the necessary data flexibly and quickly on both sides of the system without having to worry about backward compatibility and speed.

To always have a source of truth and be able to audit any events coming through the system, the application was built based on an event sourcing pattern, which due to implementation way had some known efficiency issues in cases of load from the intensive data flow. Thanks to Leobit’s senior-level architect optimizations, the experts enhanced the scalability of the jewelry platform, provided an opportunity to easily add the necessary customizations and new functionality, and most important, guaranteed optimal platform performance with an increased number of users even for such intensive operations like real-time messaging.

Our developers implemented ‘events’ functionality in very short terms allowing platform users to host, attend, and participate in digital jewelry trade shows. They considered the solution from buyers’ and suppliers’ perspectives and thought like two different categories of users. For example, buyers can find relevant sellers by providing in-system recommendations. Suppliers can track visitors’ activity in their e-booths in real-time and adjust their offers based on buyers’ preferences and internal analytics.

Profound analytic is important not only for platform users but also for business owners. Integration with Pendo service performed by Leobit’s experts helped our customer to better evaluate user experiences in their platform and prioritize which features to build next. It will make a platform even more valuable in the eyes of end-clients.

The customer was also impressed with Leobit’s speed of closing senior-level candidates for their development needs. Leobit augmented their team with 3 tech leads and 1 senior specialist within the first week of cooperation. Later on, when the necessity arose, Leobit found 5 senior engineers and 1 senior software architect in a 1-month period. The speed of reaction and the quality of our work helped Leobit to garner the customer’s trust and confidence. The upcoming plans of this partnership is the development of a mobile application, so the customers’ clients never miss a single opportunity even when they’re on the go.

Jewellery brands, professional buyers, industry suppliers and retailers looking for improved operational efficiency and enhanced business results

After the first eight months of cooperation with Leobit, the customer’s platform was ready to host a 2-day virtual trade event. The digital jewellery exhibition was visited by over 5000+ visitors. The platform’s security and reliability made attendees’ experiences seamless and provided them with the majority of possibilities available only at offline trade shows before.

The customer’s key client, Swarovski, increased user rate and vendor engagement by over 300%. It’s an impressive figure considering that all physical trade shows were cancelled due to a global pandemic.

Expanded functionality and 8x faster real-time pricing calculations increased customer value of the platform and Leobit’s partner grew its revenue by 26% in half-yearly results thanks to a subscription monetization model.