Price: 0
Number of applications: 5
25.12.25 (inclusive)
By agreement
Idea
ICT tasks
Media sphere
Neurotechnology and artificial Intelligence
Software/ IS
- High costs for cloud APIs and inefficiency of support - Lack of control over the model and dependence on external providers
1. Output documents of the study: Analytical report — comparison of 5-7 models with recommendations A working prototype is LLM, deployed locally, integrated with the knowledge base Test Report — quality, performance, and safety results Technical Document — architecture, API, deployment The implementation plan is a 3-4—month road with a budget and deadlines Executive Summary — a short summary for management with ROI 2. Positive outcomes Reducing cloud API costs by 40-60% Automation of 50%+ incoming technical support requests Increased privacy: all data remains on-premise
Pavel Andreevich Skripa
Purpose and description of task (project)
Evaluate the possibility of implementing a locally deployed LLM for: - Processing confidential corporate information (knowledge bases, documents, client data) without sending it to third-party services - Automation of technical support (FAQ, solving typical problems, routing requests) - Reducing the cost of cloud APIs - Compliance with confidentiality requirements and local legislation (Kazakhstan)
Whatsapp
