Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I’m Kurmet Aubanov, a Machine Learning Research Engineer and Applied Mathematician with 20+ years of experience. I’m focused on invariant representation learning and memory-based architectures, exploring both neural and non-neural ML systems, including backpropagation-based models. I pursue mathematically grounded ML with emphasis on interpretability, stability, and scalable GPU-based linear algebra. I build end-to-end research-to-production pipelines and explore large-scale memory systems using Ridge/Tikhonov regularization, covariance-driven features, and multi-scale memory representations. I’m open to remote opportunities worldwide or roles in Ho Chi Minh City.…Hi, I’m Kurmet Aubanov, a Machine Learning Research Engineer and Applied Mathematician with 20+ years of experience. I’m focused on invariant representation learning and memory-based architectures, exploring both neural and non-neural ML systems, including backpropagation-based models. I pursue mathematically grounded ML with emphasis on interpretability, stability, and scalable GPU-based linear algebra. I build end-to-end research-to-production pipelines and explore large-scale memory systems using Ridge/Tikhonov regularization, covariance-driven features, and multi-scale memory representations. I’m open to remote opportunities worldwide or roles in Ho Chi Minh City.

Kurmet Aubanov

Data Scientist, AI Engineer, AI Developer, +4





Hi, I’m Kurmet Aubanov, a Machine Learning Research Engineer and Applied Mathematician with 20+ years of experience. I’m focused on invariant representation learning and memory-based architectures, exploring both neural and non-neural ML systems, including backpropagation-based models. I pursue mathematically grounded ML with emphasis on interpretability, stability, and scalable GPU-based linear algebra. I build end-to-end research-to-production pipelines and explore large-scale memory systems using Ridge/Tikhonov regularization, covariance-driven features, and multi-scale memory representations. I’m open to remote opportunities worldwide or roles in Ho Chi Minh City.…Hi, I’m Kurmet Aubanov, a Machine Learning Research Engineer and Applied Mathematician with 20+ years of experience. I’m focused on invariant representation learning and memory-based architectures, exploring both neural and non-neural ML systems, including backpropagation-based models. I pursue mathematically grounded ML with emphasis on interpretability, stability, and scalable GPU-based linear algebra. I build end-to-end research-to-production pipelines and explore large-scale memory systems using Ridge/Tikhonov regularization, covariance-driven features, and multi-scale memory representations. I’m open to remote opportunities worldwide or roles in Ho Chi Minh City.

Available to hire

Hi, I’m Kurmet Aubanov, a Machine Learning Research Engineer and Applied Mathematician with 20+ years of experience. I’m focused on invariant representation learning and memory-based architectures, exploring both neural and non-neural ML systems, including backpropagation-based models.

I pursue mathematically grounded ML with emphasis on interpretability, stability, and scalable GPU-based linear algebra. I build end-to-end research-to-production pipelines and explore large-scale memory systems using Ridge/Tikhonov regularization, covariance-driven features, and multi-scale memory representations. I’m open to remote opportunities worldwide or roles in Ho Chi Minh City.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Language

English

Fluent

Russian

Fluent

Kazakh

Fluent

Work Experience

Machine Learning Research Engineer / Senior Software Engineer at Freelance / Private Clients

January 1, 2023 - Present

Developing original ML architectures and research-level implementations. Building memory-based models, invariant learning systems, and non-neural ML architectures. Implementing advanced mathematical models for analytics and prediction tasks. Providing guidance on mathematical implementations, ML architecture, and research methodologies, from prototyping to production deployment.

Lead Backend Engineer / Technical Architect at Turing Enterprises, Inc.

January 1, 2022 - December 31, 2023

Led backend development for Space Metaverse using Elixir/Phoenix. Designed scalable architecture for high-load distributed systems, and implemented CI/CD pipelines and DevOps best practices.

Director / CEO at SmartPoint LLP

January 1, 2017 - December 31, 2023

Founded and led a technology company delivering mission-critical software solutions. Built BPM systems (Elixir/Phoenix) for drilling operations, and provided infrastructure for government-level conferences. Led financial modeling and audit for fintech initiatives.

Financial Analyst / Economist at Almex Petrochemical LLP

January 1, 2014 - December 31, 2023

Developed financial-economic models for major petrochemical enterprises and built time-series forecasting models for Brent crude oil and raw material prices. Delivered market analyses and strategic insights for investment decisions.

Risk Scoring & Mathematical Models for Banking at Khalyk Savings Bank of Kazakhstan

January 1, 2009 - December 31, 2012

Developed scoring models for retail banking using statistical and ML techniques and built risk assessment models for large, medium, and small business clients. Performed big data analytics for credit risk evaluation and applied mathematical modeling to economic analysis.

Machine Learning Research Engineer / Independent Research at Independent Research

January 1, 2024 - Present

Designing and implementing memory-based ML architectures and invariant representations; exploring Ridge/Tikhonov regularization as learning primitives; feature construction pipelines and multi-scale memory systems; optimizing GPU pipelines; evaluating on reconstruction tasks such as MNIST.

Independent Researcher – Memory-based ML Architect at Independent Research

January 1, 2024 - Present

Designed and implemented memory-based ML architectures without neural networks, leveraging large-scale Ridge/Tikhonov regularization as core learning primitive. Built invariant feature representations from token windows, covariance structures, and sparse φ mappings; implemented discrete and continuous memory vocabularies with hard quantization and polynomial memory expansions across multiple degrees of freedom. Developed multi-scale memory systems with sliding windows and shared representations, enabling stable learning under stride and context shifts. Optimized GPU pipelines for large sparse and dense linear systems, including batched covariance accumulation and accelerated argmax operations. Demonstrated reconstruction improvements on MNIST and Fashion-MNIST as memory dimension scales; used regression-based and neural baselines for evaluation.

Risk Analyst / Economist at Khalyk Savings Bank of Kazakhstan

January 1, 2009 - January 1, 2012

Developed retail banking scoring models, risk assessment models for client segments, and performed big data analytics for credit risk; applied mathematical modeling to economic analysis and decision support.

Education

Applied Mathematics (Specialization: Applied Mathematics, Mathematical Modeling, Optimization Theory) at Al-Farabi Kazakh National University

January 11, 2030 - January 1, 2004

Financial Modeling at S. Toraigyrov Pavlodar State University / Ernst & Young Business Academy

January 1, 2009 - January 1, 2009

Prize-winning placements in mathematical Olympiads at Republic Physics and Mathematics School (RFMSH)

January 11, 2030 - February 11, 2026

Bachelor's degree at Al-Farabi Kazakh National University, Faculty of Mechanics and Mathematics

January 11, 2030 - January 1, 2004

Financial Modeling at Ernst & Young Business Academy

January 11, 2030 - January 1, 2009

Bachelor's degree at Al-Farabi Kazakh National University

January 11, 2030 - January 1, 2004

Certificate in Financial Modeling at Ernst & Young Business Academy

January 11, 2030 - January 1, 2009

Qualifications

Certificate in Financial Modeling

January 11, 2030 - January 1, 2009

Industry Experience

Software & Internet, Financial Services, Manufacturing, Professional Services, Education, Energy & Utilities, Government, Media & Entertainment, Other