Open to opportunities

Rauan Akylzhanov

@rauanakylzhanov

Message

I'm a Senior ML/NLP Engineer specializing in LLM production and deployment.

Kazakhstan

Message

What I'm looking for

I'm seeking a senior role building production LLM systems and real-time inference pipelines, with a focus on multi-GPU optimization, scalable MLOps, and measurable product impact within collaborative, engineering-driven teams.

I build and deploy production LLM-powered services focused on real-time reasoning, multi-GPU inference, and low-latency NLP pipelines. I design high-throughput vLLM inference stacks (RTX 4090), FastAPI-based serving architectures handling thousands of daily queries, and evaluation protocols to benchmark fine-tuned models, with measurable gains in throughput, latency and revenue.

I have improved model conversion and inference latency by porting Python algorithms to C++, led LLaMA/LangChain chatbot and CTR prediction projects on AWS, and implemented MLOps automation with Terraform, SageMaker, Kubeflow and Airflow. I hold a PhD in Mathematics and combine research rigor with production engineering to deliver scalable, reliable ML systems.

Experience

Work history, roles, and key accomplishments

Current

Senior ML/NLP Engineer

Current

HighSky

Jan 2024 - Present (1 year 7 months)

Developed and deployed high-performance LLM-powered web services for real-time reasoning, serving thousands of queries daily; optimized multi-GPU vLLM inference on RTX 4090 to maximize throughput and minimize latency. Ported semantic segmentation from Python to C++, reducing request latency from 15s to 1s and raised subnet metric from 0.72 to 0.96, improving revenue.

LLM RTX 4090 fastAPI C#Benchmarking Model Evaluation

Senior ML/NLP Engineer

Delivery Hero

Jan 2022 - Jan 2024 (2 years)

Architected and deployed an LLaMA 2-powered chatbot using LangChain, improving recommendation accuracy and raising user satisfaction by 12%. Improved conversion prediction accuracy by 20% via PyTorch transformer models and automated SageMaker resource provisioning with Terraform for LLM experiments.

LLaMA Lang Chain PyTorch Terraform Sagemaker A B Testing

ML Engineer

Kcell

Jan 2019 - Jan 2022 (3 years)

Developed churn, CLV and segmentation models and designed a scalable DataLake and MLOps pipelines, increasing model deployment efficiency by 20%. Fine-tuned TF-BERT for sentiment classification, boosting campaign response rates by 30%.

Hadoop Kubernetes Kubeflow Airflow RNN BERT Ansible

Research Associate

Imperial College London

Jan 2017 - Jan 2019 (2 years)

Taught introductory data science courses including High Performance Computing (M3C), delivering lectures and lab sessions and supporting student research projects. Contributed to academic instruction and course material development.

High Performance Computing Teaching Python C#Distributed Computing Research

Faculty Teaching Assistant

Moscow State University

Jan 2012 - Jan 2014 (2 years)

Assisted in teaching data structures, algorithms and C/C++ programming, led labs and graded assignments for undergraduate computer science courses. Supported student learning and practical programming exercises on Linux.

C#Data Structures Algorithms Linux Teaching Lab Instruction

Education

Degrees, certifications, and relevant coursework

Imperial College London

Doctor of Philosophy, Mathematics

2014 - 2018

Activities and societies: Member of the Imperial College Data Science Society and Machine Learning Society; participated in hackathons and industry engagement (2014–2017).

Conducted PhD research in non-commutative analysis at Imperial College London with EPSRC funding from 2014 to 2018.

Lomonosov Moscow State University

Specialist in Computer Science, Computer Science

2007 - 2012

Grade: With Honours

Completed a Specialist degree in Computer Science with honours, focusing on optimization theory, statistical inference, high-performance distributed computing, and numerical methods.