Applied ML Engineer | Generative AI & Multimodal Video Pipelines Specialist
I am a BBA in IT student at Maqsut Narikbayev University with a strong interest in Artificial Intelligence and Machine Learning.
My academic focus combines business, IT, and applied research, with the goal of building a solid ML engineering profile. Alongside my studies, I work in AI-driven media production, creating generative content (visual, video, audio) for television projects. This experience sparked my interest in understanding how AI models work under the hood and motivated my transition toward Machine Learning Engineering.
Black and White Production
Aug 2025 — PresentDesigns AI pipelines and uses multimodal prompt engineering to support production workflows for corporate clients and national television.
• Designing and implementing end-to-end AI-driven pipelines for corporate clients and national television. • Utilizing Multimodal Prompt Engineering to optimize production workflows and creative outputs (JSON Prompt Engineering).
International School of Economics, MNU
Sep 2024 — PresentDeveloped VBA macros for academic auditing and data processing, and worked on Random Forest weather modeling and a Gender Economics literature review.
• Assisted Prof. Stanislav Yugay by developing custom VBA macros to automate academic auditing and data processing. • Exploring predictive weather modeling using Random Forest algorithms and conducted comprehensive literature review on Gender Economics, bridging business economics with applied ML.
Maqsut Narikbayev University
2024Developed a full-stack AI platform generating photorealistic try-ons and 360° videos using Gemini 3.1 Flash Image and Veo 3.1 via Vertex AI Long Running Operations. Conducted a 20-page business and data analysis of the Kazakhstani e-commerce market, modeling Unit Economics (CAC, LTV) and financial forecasts (MRR/ARR), including a structured data pipeline: photo upload → base64 encoding → multim...
Built an AI-augmented web application chaining Natural Language Processing with geolocation data. Translated user prompts into search keywords via Gemini 2.5 Flash Lite to extract physical locations using the Maps API.
Used Pandas for data preprocessing and structuring of item attributes (weights, values). Created Matplotlib visualizations to analyze the correlation between resource volume and total value. Automated data extraction from dynamic web elements, reporting 100% accuracy in high-load environments.
Maqsut Narikbayev University
