Zhenis Yessimbekov

Founder & Full-Stack Engineer

Serik

Founded and built an AI food-analysis application with RAG, image processing, and AWS deployment; beta satisfaction exceeded 90%.

• Founded and independently built an AI-powered food analysis application from 0 to production, covering 60,000+ food products in the knowledge base. • Integrated an LLM as the core reasoning layer, enabling interpretation of food labels, nutritional analysis, and personalized dietary recommendations with 90%+ user satisfaction in beta testing. • Built a RAG pipeline with semantic vector search over a 50K+ document nutrition knowledge base, reducing irrelevant LLM responses by 70% compared to a prompt-only baseline.

Data Scientist

JSC Kazakhtelecom

Kazakhstan's national telecommunications operator.

Built production computer vision, LLM, RAG, and predictive ML systems, including pipelines processing 500K+ daily events.

• Video Analytics: Developed and deployed computer vision pipelines for real-time video stream analysis across 200+ surveillance and infrastructure cameras, achieving 92%+ object detection accuracy using YOLO-based models. • LLMs: Built and integrated large language model solutions for internal automation tasks, reducing manual document processing time by 45% and enabling natural language querying over structured telecom data.

Full-Stack Developer

ORDES.KZ (datalook.kz)

Built and deployed a sales analytics service for 100K+ daily transactions; cut data processing time by 60%.

• Architected a scalable sales analytics service from scratch, handling 100K+ daily transactions across multiple retail locations. • Implemented backend integrations and data aggregation pipelines, reducing data load and processing time by 60% through query optimization and smarter data handling logic. • Designed a normalized PostgreSQL schema supporting 10+ analytical report types with sub-second query response times on datasets of 1M+ rows.

ML Engineer (Intern)

NITEC

National Information Technologies, a Kazakhstan IT company providing information systems and e-government services.

Trained and evaluated ML models and prototyped LLM applications; contributed to integrating two AI solutions into production workflows.

• Trained and evaluated 10+ machine learning models across classification, regression, and NLP tasks, achieving average accuracy improvements of 12% over baseline benchmarks. • Researched and prototyped LLM integrations for 3 internal AI applications, reducing manual processing time by approximately 25% per workflow. • Processed and cleaned datasets of 500K–2M records using pandas and scikit-learn pipelines, improving data quality scores by 30%.

Bachelor, Software Engineering

Astana IT University