Technological tasks
The Tech-tasks module is tasks related to the development, implementation and improvement of new technologies, products and processes. If you have development, implementation and other needs you can post information below.
Industrial Tour to JSC "AK Altynalmas"
Astana Hub invites IT companies and technology teams to take part in an industrial tour at "AK Altynalmas" JSC. The participants will visit the company's production facilities, see the real processes of the gold mining enterprise and get acquainted with the current technological challenges of the industry. The date is 09/25/2026.
Decision acceptance deadline
22.09.26 (inclusive)
Preferred systems
Other solutions for the creative industryNumber of applications
21
A method for increasing the diversity of responses from large language models
To develop a portable method for increasing the diversity of responses from large language models while maintaining their quality and consistency with the user's request. The solution should work with different models and several types of tasks: from choosing options from a given set to generating ideas, titles, and short stories. The solution format is a software prototype that applies the developed method and allows you to compare the results with the original query without using the developed method.
Decision acceptance deadline
01.10.26 (inclusive)
Preferred systems
Neurotechnology and artificial IntelligenceNumber of applications
8
Development of the module "Reconciliation reports with counterparties" in the tax office
Creation of a separate section in the tax office for the formation, automatic reconciliation, sending, approving and signing reconciliation reports with counterparties using EDS. Provide integration with 1C:Accounting for the automatic exchange of data on mutual settlements and the exclusion of repeated manual entry of information.
Decision acceptance deadline
01.10.26 (inclusive)
Preferred systems
Other solutions for the creative industryNumber of applications
9
Comparison of the company's legal name and brand (matching BIN ↔ trade name)
Implement a solution for matching the legal name/BIN and the company's brand (matching BIN ↔ trade name)
Decision acceptance deadline
01.10.26 (inclusive)
Preferred systems
Information processing and transformationNumber of applications
9
AI agent to support end-to-end autotests in CI: crash analysis, test correction, and coverage expansion
To develop a portable AI agent that runs inside the continuous integration (CI) pipeline and supports end-to-end automated tests (E2E tests) on Playwright for web products. All code changes are processed by the agent as merge requests (PR/MR); they are accepted only after a review. After each deployment, the agent runs the tests on the stand and classifies the result.: 1. Tests have failed — analyzes the log, traces, and screenshots and determines the source of the failure: the test, product, or infrastructure. The outdated test corrects and opens the PR/MR. The alleged product defect is framed as a task with supporting materials and an indication of the likely location of the defect in the code. 2. Tests passed — compares coverage with product changes made after the last processed commit. If he finds an uncovered change, he writes a script, runs it on the stand and opens PR/MR. If there are no such changes, it fixes "no changes required". 3. There are no tests — it creates a set from scratch according to the approved list of scenarios. If the interface elements cannot be reliably identified, first prepares a separate PR/MR with stable selectors, and writes the tests in the next run after deployment. 4. If the reason cannot be determined, the booth is unavailable or there is not enough data, it completes the run with the status `blocked`, describes what needs to be clarified, and does not change the code. Xena Soft independently developed a working prototype and uses it on two of its projects: one pipeline on Azure DevOps Pipelines, the other on GitHub Actions. Each project has its own conventions, and configuration, access rights, and branches are configured separately. It is necessary to turn the prototype into a product: a single core, three CI adapters (GitHub Actions, GitLab CI, Azure DevOps), a panel with a history of runs and metrics, documentation according to which a new project is connected in one working day.
Decision acceptance deadline
30.09.26 (inclusive)
Preferred systems
Neurotechnology and artificial IntelligenceNumber of applications
11
Refinement of the IDEAWA mobile AI application backend: query processing architecture, recommendations, external APIs and AI integration
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.
Decision acceptance deadline
30.09.26 (inclusive)
Preferred systems
Processing and storage of Big DataNumber of applications
10
Refinement of the frontend mobile application IDEAWA
Get a stable version of frontend IDEAWA for iOS and Android, corresponding to the existing design layouts and technical specifications, without critical UI/UX errors. We are working with an existing application and an existing backend. Full application development from scratch is not required.
Decision acceptance deadline
30.09.26 (inclusive)
Preferred systems
Processing and storage of Big DataNumber of applications
8
OCR on hosting
A task for the community: local OCR of PDF scans (RU/KK) instead of Gemini What is it about In our service (verification of packages of documents for the state text is extracted from the downloaded PDF files for further analysis. Most files have a text layer and are parsed locally. But some of them come in scans without a text layer — they are now recognized by the cloud-based Google Gemini 2.5 Flash. We want to get away from the cloud and do it locally, on our hosting. What needs to be done A module that accepts a PDF scan and returns the extracted text (plain text, UTF-8), verbatim and in the correct reading order. It should work offline, without external APIs. Important caveat on the type of documents This is not the recognition of passports, identification cards, complex forms or manuscripts. Scans are ordinary printed text documents: charters, certificates, contracts, business plans, letters. That is: mostly solid printed text, 1 column; machine font (not a manuscript); there may be simple tables, a stamp/signature at the bottom — but this is not the main content. This greatly simplifies the task: you don't need the "intelligence" of the multimodal LLM level for complex layout — you need reliable OCR of printed RU/KK text. Specificity Languages: Russian + Kazakh (ә ғ қ ң ө))))), sometimes in the same document. Quality: scans of different quality (distortion, noise, DPI 150-300). Critical: numbers, amounts, BIN/IIN (12 digits), dates — an error in a digit is unacceptable. Input: PDF up to 15 MB, 1-30 pages. Output: UTF-8 text, top—to-bottom / left-to-right order; tables - line by line. Integration: A PHP module or a local CLI/HTTP service called from PHP (CodeIgniter 4 / PHP 8.2 stack). Allowed stack Everything that is hosted: Tesseract 5 (rus + kaz LSTM), PaddleOCR / EasyOCR, docTR; rasterization of PDF→image (pdftoppm / Ghostscript); preprocessing (OpenCV / ImageMagick: deskew, denoise, binarize). Combinations are welcome. Acceptance criteria Benchmark: we will provide a set of ~150-200 real impersonal PDFs (RU/KK) with a reference text. Metrics: CER / WER vs. benchmark; Field-accuracy for critical fields (BIN, amounts, dates). The goal: accuracy is no worse than Gemini 2.5 Flash on the same set (we will publish its results as a baseline). Performance: less than N seconds per document (to specify the hardware), completely local. The format of the submission Repository with code + README (installation, dependencies, launch). A benchmark run script that outputs CER/WER and field-accuracy. A brief description of the pipeline. Deadline: 01.10.26
Decision acceptance deadline
29.09.26 (inclusive)
Preferred systems
Information processing and transformationNumber of applications
8
Гарантированно не списать деньги дважды без распределённой транзакции
Разработка мобильного приложения с защищённым механизмом проведения финансовых операций, исключающим повторное списание средств при сетевых сбоях, тайм-аутах и повторной отправке запросов. Система должна корректно определять состояние операции даже при отсутствии однозначного ответа от внешнего платёжного провайдера.
Decision acceptance deadline
17.09.26 (inclusive)
Preferred systems
FinTechNumber of applications
3
Pitching of technological solutions in the field of Industry 4.0 at AI & Digital Bridge 2026 with a prize fund of 5 million tenge
Search and selection of technological solutions in the field of Industry 4.0. Participants will have the opportunity to present their digital and AI solutions to large enterprises of the oil and gas and mining industries at the AI & Digital Bridge 2026 site. Reflect in the presentation: the company/team, the problem and its solution, the product description, the target audience, existing analogues and advantages, roadmap, business model and additional important information.
Decision acceptance deadline
21.09.26 (inclusive)
Preferred systems
Цифровизация, автоматизация и ИИ в промышленностиNumber of applications
57
