Building end-to-end machine learning systems — from data pipelines and model training to APIs, evaluation, and deployment.
I'm an aspiring Machine Learning Engineer based in Astana, Kazakhstan, with a background in systems, networking, and infrastructure. I focus on building practical ML projects that go beyond notebooks: reproducible pipelines, model evaluation, inference services, monitoring, and deployment-oriented engineering.
My current interests include Computer Vision, recommender systems, classical ML, Generative AI, and ML systems.
ML & Data
Python · PyTorch · scikit-learn · LightGBM · Hugging Face · Diffusers
Computer Vision & GenAI
YOLO · VLMs · LoRA / PEFT · Video Data Pipelines
ML Engineering
FastAPI · MLflow · Docker · Ray · Accelerate · DeepSpeed · CUDA
Systems
Linux · Bash · Git · CI/CD · SQL · Networking
Computer Vision · Object Detection · ML Service
End-to-end road-damage inspection project built around object detection, with training/evaluation workflows and deployment-oriented inference.
Classical ML · LightGBM · FastAPI · MLflow
Production-oriented regression project for taxi trip-duration prediction, covering feature engineering, experiment tracking, model serving, and monitoring-oriented workflows.
Recommender Systems · Ranking · Implicit Feedback
Two-stage recommendation and ranking system built from e-commerce interaction data, with temporal evaluation and service-oriented project structure.
Generative AI · Video · VLM · Distributed ML
Research-oriented pipeline for converting raw video into curated, captioned datasets for generative-model experiments, including filtering, VLM captioning, distributed preprocessing, GPU benchmarks, and diffusion LoRA experimentation.
- Practical experience building end-to-end portfolio ML systems, not only training notebooks.
- Strong foundation in Linux, networking, virtualization, and infrastructure from previous systems work.
- Experience experimenting with GPU training, distributed workloads, model serving, and ML tooling.
- A systems-oriented approach to ML: reproducibility, observability, deployment, and maintainability matter alongside model quality.
I'm currently strengthening my ML engineering portfolio and looking for opportunities where I can contribute to real-world Machine Learning, Computer Vision, or applied AI systems.
Machine learning is most useful when the model becomes a reliable system.