About:
Darya is an AI/ML engineer with deep specialization in computer vision, natural language processing, large language models (LLMs), and generative AI. With end-to-end experience across the machine learning lifecycle, she brings a rare combination of mathematical rigor, advanced engineering skill, and real-world problem-solving ability to every project she leads.
Darya has successfully built and deployed production-grade AI systems for retail, trading, and enterprise automation domains. She has fine-tuned multimodal foundation models like Qwen2.5-VL for visual-language tasks, engineered retrieval-augmented generation (RAG) pipelines for intelligent document systems, and delivered real-time video analytics solutions using state-of-the-art detection and OCR frameworks.
Her technical versatility spans modern MLOps tooling, synthetic data generation, and scalable model deployment with Ray Serve and Docker. With a passion for experimentation and a strong foundation in mathematics and research, Darya continuously drives innovation at the intersection of AI, engineering, and business.
Tech expertise:
- Fine-tuning and deploying multimodal LLMs (Qwen2.5-VL, LoRA, INT8 quantization)
- RAG, prompt engineering, and LangChain pipelines
- Computer vision and OCR systems (YOLOv11x, PaddleOCR, OpenCV)
- Scalable ML infrastructure (Ray Serve, FastAPI, Flask, Docker)
- MLOps lifecycle automation (Apache Airflow, MLflow, Blender)
- Synthetic data generation and image augmentation for low-data scenarios
- Video analytics and object recognition pipelines with caching and stream handling
- Deep applied mathematics for algorithm design and model optimization