Skip to content
AA

Alsu Akhatova

@alsu1311Connected with DSML. This person's Telegram account has been confirmed through DSML Kazakhstan.

ML Engineer | Python, PyTorch, SQL

ML Engineer with a background in data analytics and quantum technologies. Focused on Computer Vision, with hands-on experience in Python, PyTorch, scikit-learn, and deep learning. Completed ML/CV/NLP programs from MIPT DLS, MSU, and Andrew Ng.

Open to

Joining pet projectsJoining competitions & hackathonsPaper discussionsStartup collaboration

Career

Experience: 4 years

Experience

Analyst

RESO-Garantia

A major Russian insurance company offering auto, property, health, and life insurance to individuals and businesses.

Sep 2023 — Present

Built SQL validation and claim-processing tools for insurance analytics, including a script that made checking 1.5 times faster.

• Developed an SQL conversion script to validate an ML model for predicting base rates, enabling prediction accuracy assessment and further algorithm improvements. • Automated processing of a large volume of insurance claims with a script that made checking 1.5 times faster and reduced manual work. • Fixed a region mapping error, ensuring correct calculation of territorial coefficients and improving pricing accuracy. • Developed SQL scripts, functions, and procedures for analytics, and maintained regular table updates to keep reporting data current.

Research Intern

Keldysh Institute of Applied Mathematics, Russian Academy of Sciences

A Russian Academy of Sciences research institute focused on applied mathematics and computational methods.

Feb 2021 — Jun 2022

Modeled laser radiation propagation in Python and numerically found an absorption-regime transition point that was unavailable analytically.

• Developed a Python numerical model of laser radiation propagation, implementing explicit and implicit Euler methods to solve a system of differential equations. Results agreed well with experimental data. • Numerically determined the transition point between laser radiation absorption regimes, which could not be found analytically.

Education

BSc Applied Physics and Mathematics

Moscow Institute of Physics and Technology

2018 — 2022

Selected projects

CLIP-Guided Domain Adaptation of Image Generators

Reimplemented StyleGAN-NADA, reproducing its training process and partial generator-layer unfreezing, then fine-tuned the model for five text prompts: Elf, Witcher, Pixar, Mona Lisa, and Sketch.

Open project

Achievements

Semi-finalist, All-Russian Student Olympiad 'I Am a Professional', Quantum Technologies track

All-Russian Student Olympiad 'I Am a Professional' · 2022

Winner, Best Bachelor's Presentation nomination

KMU ITMO Conference

Skills

PythonSQLPyTorchscikit-learnPandasNumPySciPyMatplotlibMachine LearningComputer VisionNLPProbability TheoryAlgorithms and Data StructuresGit