People in DSML Kazakhstan

My research interests lie from developing large multimodal models to practical computer vision tasks. Prior to that I focused on developing methods for image-to-image tasks such as super-resolution, image/video enhancement, and digital image processing part of which have been accepted to top-tier conferences and deployed in production environments serving millions of users.

Software QA Engineer with 8+ years of experience testing web, desktop, API, and embedded automotive EV systems, combining manual and automation approaches to improve release quality and reduce production defects, along with a growing focus on using AI and Data to improve testing effectiveness. Experience spans REST microservices, trading platforms, banking, TV streaming, e‑commerce, and EV wall-box chargers in agile and hybrid (on-site, remote) environments. Quickly adapting to new tech stacks, owning automation frameworks, and collaborating closely with developers, testers, and product stakeholders. Using AI tools at a basic level to speed up test design, log analysis, and documentation, actively developing Data skills to move towards QA + Data hybrid roles.

IT Park Uzbekistan — лучшая площадка для развития IT-бизнеса, инноваций и экспорта цифровых услуг. Объединяем технологии, талант и возможности, создавая благоприятную экосистему для стартапов, международных компаний и инвесторов.

Computer Vision & Machine Learning from Kyrgyzstan. Mostly experienced in real-time object detection, instance/semantic segmentation, classification, pose estimation and tracking. Open-source developer

Final-year Information Systems student at SDU University focused on data analytics, SQL, Python, and business intelligence. Interested in transforming raw data into clear insights, dashboards, and practical recommendations for business decision-making. Looking for a Junior Data Analyst or Data Analyst Intern role to apply analytical skills to real-world business problem

AI/LLM engineer from Kostanay. I build production RAG systems, LLM agents and automation pipelines — and I measure them instead of guessing. Examples: GraphRAG against vanilla RAG on the same corpus (+20 points on multi-hop questions), a document pipeline turning a 116-page tender into 745 requirements with page references and confidence, recall 92.2% on a hand-labeled gold set, and a deterministic evaluation harness with no LLM-as-judge (accuracy 85 to 100%). Stack: Python, FastAPI, LangGraph, Qdrant, Neo4j, PostgreSQL, Docker, Kubernetes. Open to remote work and contract projects.