LLM research / AI engineering
Aleksandr Medvedev · Innopolis University
Third-year Applied AI student / seeking internships
I research LLM quality and build auditable AI systems, from held-out evaluation and RAG to automated model deployment.
I’m a third-year Applied Artificial Intelligence student at Innopolis University (BSc, 2024–2028). I study LLM quality through held-out evaluation and build auditable services with RAG and human review. I have led a research team and shipped reproducible ML pipelines.
Python · PyTorch · TensorFlow · Transformers · FAISS
Held-out evaluation · bootstrap confidence intervals · RAG
FastAPI · PostgreSQL · DVC · MLflow · Docker
LLM research / routing / evaluation
Compared five Qwen2.5 routing strategies over 1,000 runs on 500 MMLU questions, using held-out evaluation and bootstrap confidence intervals. Built traceable RAG, abstention, and human review into the service.
Research leadership / 3D search
Led a four-person research team through a 150-query evaluation across five scenes, with ScanNet and Replica pilots for natural-language 3D search.
MLOps / training / serving
Automated DVC, MLflow, FastAPI, and Streamlit delivery with 10 CI tests. The held-out model reached R² 0.993 and RMSE 1.28 on five unseen hulls.