Skip to content
View AlgazinovAleksandr's full-sized avatar
🎯
Focusing
🎯
Focusing
22 followers · 48 following

Block or report AlgazinovAleksandr

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
AlgazinovAleksandr/README.md
Hi, I'm Aleksandr! Typing SVG

Portfolio LinkedIn Google Scholar Email Telegram

About

I'm an AI Engineer at Sberbank (the largest bank in Russia), working on Multi-Agent Systems, and a Computer Science Master's student at Tsinghua University. I have 4+ years of ML experience spanning industry and research, with a background across NLP, LLMs, Multimodal AI, Time Series, MLOps, and Classic ML.

Currently working on: production-ready agentic pipelines (LLMs, structured outputs, multi-agent orchestration) and multimodal AI research.

Philosophy

The role of a Data Scientist has fundamentally shifted. Today it means being an AI Engineer — someone who designs end-to-end systems and pipelines, not just calls .fit() and .predict().

Keeping up in this field demands rapid learning, constant adaptation to new tools, and applying knowledge on real data. Understanding why a method works, not just that it works, is what separates solid engineers from the rest. That's why I consistently work on pet projects, read the recent papers, and study new tools.

Research
MATE Multi-agent system for context-aware modality conversions · GitHub
TRACE LLM fine-tuned for transparency-focused reliability scoring of web content · GitHub
Springer Book Chapter Embedding AI into network devices to improve efficiency, latency, and topology optimization
ODS AI Talk Anomaly Scoring for Preventive Detection of Failures in Information Systems

Stack

AI & LLMs

Python PyTorch HuggingFace LangChain LangGraph AutoGen

Classic ML & Data

scikit-learn CatBoost LightGBM NumPy Pandas PySpark

Engineering & Infra

FastAPI Docker Kubernetes Airflow PostgreSQL Git Bash

Find Me
Portfolio & CV algazinovaleksandr.github.io
Email algazinovalexandr@gmail.com
Telegram @krasnorechivyy
Blog (in Russian) My Amazing Channel

Pinned Loading

  1. STREAM-RL STREAM-RL Public

    Using Reinforcement Learning for mitigating vehicle traffic by providing user instructions

    Jupyter Notebook 13 1

  2. Option-Pricing Option-Pricing Public

    We solve the problem of option pricing using Machine Learning and Deep Learning methods. The materials presented in this repository are prepared for my Bachelor's Thesis

    Jupyter Notebook 9

  3. Financial-Analysis Financial-Analysis Public

    Solving financial problems and analyzing financial data with Data Science tools (portfolio analysis, stock price dynamics trends analysis, etc.)

    Jupyter Notebook 5

  4. NLP-LLMs-Fine-Tuning NLP-LLMs-Fine-Tuning Public

    Solve various NLP tasks by applying modern NLP architectures and tooling - from foundational embeddings to transformer‑based models like BERT and GPT; Use methodologies like fine‑tuning and prompt …

    Jupyter Notebook 8 1

  5. Time-Series-forecasting Time-Series-forecasting Public

    Multivariate time series analysis and forecasting. Use multiple models, such as CatBoost, Prophet, LSTM, Seq2Seq, Transformer, and AutoML to compare classical, ml-based, and deep learning approaches

    Jupyter Notebook 6

  6. Multi-Agent-MATE Multi-Agent-MATE Public

    The first open-source and lightweight multi-agent system for comprehensive modality conversion tasks

    Jupyter Notebook 14 1