Arsen Abzhanov

@arsenio

Data Scientist | Generative AI and Computer Vision

I am an AI engineer. I have experience in making RAG chatbots, AI agents, research. I have an interest in explainable AI.

Career

Experience: 1 years

Experience

Middle Data Scientist

Halyk Bank

Dec 2025 — Present

Built a natural-language-to-SQL AI agent and maintained RAG chatbots.

• Developed an AI agent that translates natural language queries into SQL queries. • Performed maintenance on RAG chatbots in Dify. • Stack: Python, FastAPI, LangChain, LangGraph, Docker.

AI Engineer Intern / Junior Automation Engineer

Home Credit Bank

Jul 2025 — Nov 2025

Automated email and attachment processing, masking sensitive data and building a deposit-offer monitoring and RAG chat platform.

• Developed an AI-driven system to automate email and attachment processing. • Secured sensitive data with automated masking before LLM integration. • Streamlined operations by enabling context-aware responses and automatic task assignment. • Built an end-to-end deposit offer monitoring and RAG chat platform for top banks in Kazakhstan. • Stack: Python, MongoDB, FastAPI, Hugging Face Transformers, LangChain, LangGraph, Gemini, spaCy, Docker.

Cloud Engineer Intern

Bespin Global

May 2024 — Jun 2024

Configured cloud networks and virtual machines, deployed a RAG chatbot, and used Azure face recognition.

• Set up and configured networks, including VPNs and virtual machines. • Deployed a chatbot that uses Retrieval Augmented Generation. • Used Azure face recognition. • Stack: Azure.

Software Engineer

Trilogy (Crossover)

Aug 2022 — Oct 2022

Built and deployed web applications, including one using OpenAI APIs, and pitched project ideas to management.

• Developed a web application with authentication and authorization, and deployed it to AWS. • Implemented a TikTok-like web application that uses OpenAI APIs. • Collaborated with colleagues to develop and pitch project ideas to management. • Stack: Django, TypeScript, AWS, NoSQL, Python.

Education

Master of Science, Computer Vision

Mohamed bin Zayed University of Artificial Intelligence

2023 — 2025

Bachelor of Science, Computer Science

Nazarbayev University

2017 — 2021

Selected projects

Integration of a diffusion transformer for explainable medical classification

Improved the Explicd multimodal explainable diagnosis framework by incorporating a diffusion transformer, making diagnoses more robust to clinically irrelevant features and improving F1-score across skin lesion and breast ultrasound datasets.

Classification and Segmentation of Intrapartum Scans

Improved the MICCAI 2024 IUGC organizers' baseline multitask model, increasing classification accuracy by 14% and Dice score by 6%. Earned 3rd place in the challenge.

Vision-Language Model for Explainable Diagnosis Using Chest X-rays

Improved the Xplainer explainable diagnosis framework through prompt engineering and increased its AUC score.

Drawing with Eye Tracking

Developed a Python script for real-time gaze estimation using MediaPipe, sending data to a C# script in Unity to control an interactive object and let the user draw.