AI Engineer · RAG · LLM Systems · AI Agents
I build and evaluate practical AI systems — from retrieval and context engineering
to LLM workflows, agent collaboration, fine-tuning, and applied experimentation.
- RAG & Context Engineering — retrieval quality, hybrid search, reranking, grounded generation, evaluation, and validation.
- LLM systems — inference, fine-tuning, LoRA / QLoRA / PEFT, quantization, and model evaluation.
- AI agents — tool-using workflows, durable context, handoffs, human-in-the-loop development, and GitHub-native coordination.
- Prompt Engineering — structured prompting for text, image, and video generation, including agent-oriented patterns.
|
Production-oriented RAG research project covering dense + sparse retrieval, hybrid fusion, Context Engineering, grounded generation, LLM-as-a-Judge evaluation, and confirmation / holdout validation. Focus: RAG · Qdrant · Hybrid Retrieval · Context Engineering · Evaluation |
A GitHub-native standard for passing durable project context between AI coding agents, human maintainers, and human-supervised agents. Focus: AI Agents · GitHub Workflows · Agent Memory · Human-in-the-loop |
|
Practical notebooks for LLM fine-tuning with Prompt Tuning, LoRA, and QLoRA, Whisper fine-tuning for Russian ASR, post-training quantization, inference, and measurable evaluation. Focus: LLM Fine-tuning · PEFT · ASR · Quantization · Evaluation |
A portfolio of structured prompting work for text, image, and video generation, including reasoning, refinement, ReAct, tool-selection, and multimodal prompting patterns. Focus: Prompt Engineering · Multimodal AI · ReAct · Tool Use |
I'm especially interested in building reliable AI systems that can be measured, reproduced, and improved — not just demos. My current work is centered on retrieval quality, context construction, grounded generation, AI-agent workflows, and evaluation-driven iteration.
Open to interesting AI engineering, RAG, LLM, and agent-system collaborations.