Skip to content
DSML.KZ
MB

Maksat Buranbaev

@MaksatConnected with DSML. This person's Telegram account has been confirmed through DSML Kazakhstan.

ML Engineer

I am a recent university graduate currently relocating from Russia to Astana, Kazakhstan, and I am open to job or internship opportunities. I specialize as an ML/AI Engineer. Outside of AI, I enjoy computer and card games, as well as playing Go

Open to

Learning with othersJoining pet projectsFinding a mentorHiring & referralsJoining competitions & hackathons

Career

Experience: 0 years

Experience

Data Scientist / ML Developer (Internship)

SberTech

Jun 2025 — Jul 2025

Built a natural-language SQL assistant using GigaChat and PostgreSQL; query validation and error handling reportedly reduced analytics turnaround severalfold.

• Developed an intelligent SQL assistant pipeline using the GigaChat API to automate natural-language data retrieval from relational databases (PostgreSQL). • Implemented syntax validation for generated queries and end-to-end handling of generation errors, reducing the time needed to obtain analytics severalfold. • Technology stack: Python, GigaChat API, SQL, PostgreSQL, LLMs.

Database Developer

ZAO Bankovskie Informatsionnye Sistemy

Dec 2024 — Mar 2025

Maintained and refactored banking-system database modules, improved report generation speed by 15%, and scripted cleanup of anomalous data after a system failure.

• Maintained and refactored legacy modules of an automated banking system using Progress OpenEdge (ABL). • Optimized queries and stored procedures, accelerating regular report generation by 15%. • Developed an ad hoc script to clean product databases of anomalous records after a system failure. • Technology stack: Progress OpenEdge, ABL, database optimization, refactoring.

Education

Bachelor's degree, Fundamental Informatics and Information Technologies

Omsk State Technical University

2026

Selected projects

RAG System: Intelligent Assistant for Databases

Built a Retrieval-Augmented Generation pipeline intended to answer technical questions based strictly on specialist literature. Used Sentence-Transformers to embed text chunks and FAISS to store and search indexes. The system returns answers with references to specific book sections and is described as filtering out fabricated facts.

Web Application for Go with AI (KataGo)

Developed a web interface with a FastAPI backend that communicates with the KataGo engine using the GTP protocol. Implemented real-time WebSocket communication, current-position analysis, and automatic score calculation.

Skills

PythonSQLC#ABLPyTorchTensorFlowKerasScikit-learnXGBoostCatBoostLightGBMEnsemble learningTime seriesLangChainOllamaRAGFAISSSentence-TransformersNLTKGigaChat API