About me
Konstantin Kolomeitsev is a Chief AI Officer (CAIO), AI/LLM architect, founder of Loqira Labs, and applied AI researcher based in Almaty, Kazakhstan. He helps companies assess, design, and implement AI systems: LLMs, RAG, AI agents, local models, and business-process automation. He also provides AI consulting and corporate AI training.
This page collects the facts: current roles, work history, education, and the profiles where I publish.
- Role
- Chief AI Officer (CAIO)
- Company
- Sergek Development, Almaty
- Lab
- Loqira Labs (founder)
- Location
- Almaty, Kazakhstan
- Focus
- LLMs, RAG, AI agents, local models
- Education
- M. Auezov South Kazakhstan Research University, Software Engineering (2010–2014)
From engineering to AI strategy
I have worked in software development since 2013: websites, high-load microservices, billing, IAM platforms, and distributed-system architecture. I walked the whole engineering path from developer to solutions architect.
Since 2021 I have worked at Sergek Development in Almaty: led a development team, owned application architecture, and in November 2025 became Chief AI Officer. I own the AI strategy and its execution: prioritization, the data platform, quality standards, and governance.
I founded Loqira Labs, an independent edge-AI lab: capable models for everyday tasks that run on ordinary hardware. Its first shipped product is Levitan, on-device voice input for Windows and macOS.
My research covers applied AI: knowledge transfer into small models (LLM Modules, arXiv:2502.08213), the MPT-VC neural video codec built on original operators, and long-context efficiency (Context Merging). I publish the work openly, with papers, weights, and code.
Leadership plus hands-on execution
Progressively broader scope: from engineering to platform and AI strategy ownership.
- Own end-to-end AI strategy and execution aligned with core business objectives.
- Design and deploy AI/ML systems (LLM, NLP, CV); own data governance and pipeline quality.
- Build and lead cross-functional AI teams; define standards and hiring strategy.
- Define Responsible AI: privacy, monitoring, bias mitigation, and compliance.
- Application architecture and key technical decisions.
- Team leadership, task allocation, and refactoring of existing services.
- Development of new services and integration of existing systems.
- Architecture leadership and selection of technical solutions.
- Refactoring and delivery of new services.
- Planning, prioritization, and technical mentorship.
- High-load microservices (300–800 pkt/s): TypeScript, Node.js, NestJS, PostgreSQL, React.
- Deployment automation (docker swarm) and system-integration workflows.
- Engineering standards: tests, linting, CI, and faster delivery.
- Microservices: Node.js, Express, MongoDB, PostgreSQL, GraphQL, React.
- Billing system development.
- Central authorization and IAM services: architecture, backend, and frontend.
- Shared server and client packages; Kubernetes deployment scripts.
Earlier experience (2013–2018)
- Web apps: Node.js + Express backend, React frontend.
- A blockchain platform on Node.js and Ethereum smart contracts (Solidity).
- PWA, GraphQL, React Native, Electron.
- Admin UI on React for an in-house CMS.
- A mobile app on Cordova / Ionic.
- Support and improvements for ecco.kz.
- UI customization in JavaScript.
- Development and support for biosfera.kz.
- SQL query optimization.
- Development and support for evrika.com.
- Customization of Bitrix CMS modules.
- Websites, including government sites.
- Joomla, WordPress, Drupal; PHP modules and SQL improvements.
Education
Languages
Press, interviews, and articles
External coverage and my own articles. Most of them are in Russian.
Where I publish
The same person across all platforms: research, code, models, and articles.