Price: 0
Number of applications: 10
30.09.26 (inclusive)
Monetary remuneration under the contract for the performance of an agreed amount of backend work. The possibility of further cooperation on the IDEAWA project.
MVP
ICT tasks
Media sphere
Processing and storage of Big Data
Mobile app
The existing backend logic needs to be brought to a single stable architecture for processing user requests. The system should exclude situations in which: one user's message is processed multiple times; the request is lost when the backend fails; external APIs are called unnecessarily; AI independently determines data sources or comes up with recommendations.; the client receives raw data from third-party APIs; responses from different sources have an incompatible structure; if one external service fails, the entire user script breaks down.; it is impossible to determine at what stage the error occurred.; The user is shown a response that has not yet been saved in the system. According to the IDEAW architecture, the backend should be the main control layer: it defines the query scenario, receives the actual data, filters and ranks it, and AI is used as a layer for presenting already prepared recommendations.
IDEAWA's stable and scalable backend provides a full cycle of personal recommendation generation — from receiving a user's request to issuing ready-made cards and an AI response. Expected result: improving the stability of the application; reducing the number of hung and lost requests; avoiding message re-processing; reducing unnecessary calls to external APIs; reducing the number of incorrect AI responses; faster processing due to caching and parallel requests; the ability to fully diagnose errors; the backend's willingness to increase user load; the ability to connect additional sources, categories, and personalization algorithms without redesigning the entire architecture.
Boldyshevsky Dmitry Ilyich
Purpose and description of task (project)
To develop and stabilize the IDEAWA backend pipeline, which provides a full cycle of user request processing: message → save → queue → intent definition → constraints formation → source selection → external data acquisition → normalization → filtering → ranking → AI context formation → AI challenge → validation → save → delivery to the client. The system must be: fault-tolerant; idempotent; scalable; deterministic in business logic; suitable for the subsequent expansion of the number of categories and data sources.