Salamat Quantaiuly

@salamat-quantaiulyConnected with DSML

Data Scientist

I work with data, from industrial device logs to customer reviews and machine learning projects. Recently I dug into 431,000 Flip.kz reviews to find product issues that did not show up often but mattered.

Open to

Research collaborationLearning with othersCommunity initiatives

Career

Experience: 3 years

Experience

Data Analyst Trainee

Kaspi.kz

Kazakhstan's consumer finance and technology company, known for its payments, marketplace, and fintech services.

Dec 2025 — Feb 2026

Built call performance dashboards and alerts for supervisors, and optimized reporting pipelines to cut query execution time by 50%.

• Owned call-performance analytics for operational supervisors, from data preparation through monitoring and reporting. • Replaced fragmented Excel tracking with real-time reach-rate and coverage dashboards and threshold alerts, enabling supervisors to spot dialer anomalies and reallocate call flows mid-shift. • Optimized Airflow and StarRocks reporting pipelines for client interactions and call outcomes, cutting query execution time by 50%.

Backend and Data Analytics

LLC Altyn Shyghys

Sunflower oil extraction and refining plant.

Jul 2023 — Dec 2025

Built industrial data collection and SCADA integrations, with alerts for sensor anomalies to help maintenance staff respond faster and reduce downtime.

• Developed an application for automated data collection and logging from OPC, PLC, and Modbus devices. • Built a SCADA integration wrapper enabling data logging for unsupported devices. • Used NATS and RabbitMQ transport, Python and Go microservices, and Grafana alerts to notify maintenance staff of sensor anomalies, reducing response time and equipment downtime.

Teaching Assistant

Kazakh-British Technical University

Kazakh-British Technical University is a university in Almaty with a focus on technical and business education.

Feb 2024 — May 2024

Supported an undergraduate Introduction to Machine Learning course by helping students with assignments, explaining concepts, and evaluating lab work.

• Assisted in teaching the Introduction to Machine Learning course for undergraduate students. • Supported students with assignments, explained core ML concepts and Python libraries, and facilitated Q&A sessions. • Evaluated laboratory assignments and projects, providing detailed constructive feedback.

Education

Data Factory: Data Analysis Intensive

Kaspi Lab

2025 — 2025

MSc, Data Science

Kazakh-British Technical University

2023 — 2025

BS, Information Systems

East Kazakhstan Technical University

2019 — 2023

Selected projects

Credit Risk Model: IFRS 9 PD Prediction & Calibration

Benchmarked LR + WoE and LightGBM pipelines for credit default prediction under IFRS 9. Used out-of-fold cross-validation, Platt Scaling, and quantile binning. Reached a Gini of 0.5336, KS of 0.4425, and Brier Score of 0.0409.

Open project

MuseumSCAT: Specimen Label Transcription with VLM Selective Prediction

Built a 4-bit Qwen2.5-VL-3B pipeline to transcribe specimen label fields for the Natural History Museum of Denmark. Automated cropping, structured JSON output, and confidence scoring. Reached an AURC of 0.11853 through selective prediction.

Open project

Flip.kz Voice of Customer (VoC) & RCA Study

Processed 431K customer reviews across 28K Flip.kz products, built a multi-label complaint categorization model, and developed an impact score to identify high-impact product usability issues.

Open project

Educational Figure Generation using Text Perceptual Loss

Collected a specialized textbook figure dataset, developed an autoencoder with Text Perceptual Loss, and trained a latent diffusion model from scratch. Reduced FID by 51% and TPL score by 84% compared with a standard KL-VAE.

Open project

Skills

PythonSQLETL/ELTData PipelinesData CleaningData Warehousing (DWH)Data Mart DevelopmentApache AirflowApache KafkaHadoopFastAPIStreamlitTelegram Bot APIREST APIsModel ServingModel Inference APIsPyTorchKerasMachine LearningGo