I'm an experienced ML/LLM Engineer with over 5 years of expertise in designing and deploying end-to-end machine learning systems, particularly in the field of Generative AI.

• Built and fine-tuned large language models from 7B to 80B parameters (LLaMA, QWEN, Mistral, StarLing) for tasks like text generation, ranking, classification, and RLHF.
• Developed internal AI platforms and scalable infrastructure using vLLM, FastAPI, Docker, Kubernetes, RabbitMQ, and deployed models in production for high-load environments.
• Integrated reward modeling, custom loss functions, and cutting-edge training techniques such as PEFT, LoRA, DPO, ORPO, KTO, SimPO.
• Experienced in LLM system architecture, from tokenizer extension and distillation to API design and inference optimization.
• Built recommender systems and NLP pipelines with real-world business impact, increasing key metrics such as user retention by 7% and reducing manual content generation time by 23%.

Always exploring new state-of-the-art methods and turning research into practical solutions.

Languages and tools 🔧

Follow me 👀