Skip to content
RO

Rysbek Orazymbetov

@claude

ML/NLP Engineer (Middle) | 1.5 years of commercial experience | Remote work

Hi everyone! I'm a software engineer passionate about Deep Learning and LegalTech. I build AI-driven solutions to automate legal workflows, and bridge the gap between complex neural networks and real-world legal challenges.

Career

Experience: 2 years

Experience

ML Engineer

Mindcode

Financial and legal consulting, including audit and management accounting, as well as business and legal services.

Dec 2025 — Aug 2026

Led legal LLM development and adaptation, designed document-processing and risk-detection systems, and mentored engineers and researchers.

• Led R&D on LLMs for LegalTech systems, including development, fine-tuning, and alignment of large language models for autonomous legal agents, contract analysis systems, and predictive legal analysis. • Applied Instruction Tuning and DPO to adapt models to the legal domain, focusing on reducing hallucinations, following instructions, and data security. • Designed generative systems for end-to-end legal document processing, combining OCR and computer vision for scanned contracts, seals, and signatures with deep text processing using NLP.

ML/NLP Engineer (Middle)

Ptolemay LLC

Software development company listed in the information technology, systems integration, and internet sector.

Jul 2024 — Dec 2025

Built AI solutions for fraud detection and e-commerce recommendations, with work spanning data pipelines, NLP, RAG, deployment, and client delivery.

• Developed and deployed end-to-end AI solutions, from prototype to production, for fintech fraud detection and personalized e-commerce recommendations. • Designed and optimized high-load ETL pipelines for cleaning, normalization, and sharding to prepare large datasets for model training. • Adapted and fine-tuned BERT, GPT, and T5 models for text classification, named entity recognition, and answer generation. • Evaluated model quality using custom metrics and validation. • Developed RAG systems using FAISS, ChromaDB, and Qdrant to improve answer accuracy and relevance.

Education

Bachelor's degree, Software Engineering

Astana IT University

2027

Selected projects

Fine-tuning Gemma 9B in Kazakh

Fine-tuned Gemma 9B in Kazakh using a self-created dataset and evaluated quality with LLM-as-a-Judge.

Skills

PythonPostgreSQLBashMATLABDockerGitKubernetesExcelREST APIClickHouseMongoDBGrafanaRabbitMQAirflowpandasPyTorchNumPyRiverPlaywrightSelenium