Skip to content
BG

Bekarys Gabdrakhimov

@bekarys

Senior Data Scientist

Senior Data Scientist with 8+ years of hands-on experience in building models, data processing, and visualization, as well as programming languages, including Python and SQL. I enjoy working in a team and communicating datadriven results using various tools.

Career

Experience: 8 years

Experience

Senior Data Scientist

Business&Technology Services

Jul 2023 — Present

Built production ML systems, including a service-desk RAG system and a road-quality analytics platform; an optimization model increased chrome recovery by 6%.

• Developed and deployed an LLM-powered RAG system for service desk knowledge retrieval and question answering. • Developed a CV/ML-based optimization model that increased chrome recovery by 6%. • Designed and implemented a road quality analytics platform using haul truck telemetry and suspension sensor data. • Built and maintained production-grade ML pipelines using Airflow, Docker, Amazon S3, and GitLab CI/CD.

Data Scientist (NLP)

KazDream

Apr 2023 — Jul 2023

Worked on Kazakh- and Russian-language ASR; reports an 8% improvement in the existing model’s word error rate.

• Improved word error rate of existing ASR model by 8%. • Data preprocessing by applying num2word model, removing punctuations and using noise removal model. • Training Language Model for fixing grammatical errors.

Data Scientist

ROCKWOOL

Global manufacturer of stone wool insulation and related products.

Feb 2021 — Apr 2023

Applied NLP to marketing data and developed a smoke-detection system deployed on Jetson Xavier for real-time inference.

• Performing EDA and applying NLP techniques to social media ads. • Built tools for lead generation and behavior prediction based on marketing text and engagement signals. • Built smoke detection algorithm based on dark channel prior and optical flow. • Successfully deployed on Jetson Xavier with cloud integration for real-time inference with TensorRT.

Computer Vision Specialist

ALARIS

Mar 2018 — Dec 2020

Worked on a LiDAR-based mapping robot, covering sensor data, image transmission, hardware/software integration, and CNN object classification.

• Processing LIDAR sensor data. • Image data transmission between nodes. • Hardware and software integration. • Development of CNN for classifying objects.

Education

MSc, AI and Automation Engineering

Technical University of Denmark

2019

BEng, Electrical and Computer Engineering

Nazarbayev University

2015

Selected projects

End-to-end ML project on classification of EEG signal of depressed patients

Preprocessed EEG signals using filters and ICA, and developed ML and DL models for depression diagnosis following MLOps practices.

Asian restaurant location recommendation in the city of Nur-Sultan

Collected food venue data using Geopy and the Foursquare API, then analyzed and visualized venues by frequency, density, and distance.

Blood bank management system database

Designed conceptual and logical database models, normalized to third normal form, and implemented the database with functions, triggers, events, and procedures.

Achievements

Winner, 9 km bi-liquid category, European Rocketry Challenge (with DanSTAR team)

European Rocketry Challenge

Skills

PythonSQLMachine LearningData VisualizationGitLab CI/CDPower BIPySparkCloudGitGoogle AnalyticsAzureElasticSearch