Founding Machine Learning Engineer
Nace AI
Backed by General Catalyst; $5 million raised
Oct 2024 — PresentBuilt LLM evaluation, audit-agent, model-training, safety-classification, and distributed-training systems; reported outcomes include a 90% cost reduction, 11% model improvement, and processing 50K+ annual reports.
• Built an LLM-as-judge system to evaluate AI audit agents against expert ground truth, enabling systematic quality measurement and faster iteration on agent accuracy.
• Developed a standalone evaluation toolkit that scores AI-generated audit workpapers offline, removing dependency on the full production stack and accelerating agent debugging.
Machine Learning Engineer
Otbasy365 LLP
FinTech company for small businesses
Oct 2023 — Sep 2024Built data pipelines, an ML scoring system, and customer-classification ensembles; specific performance results are not provided.
• Built a data pipeline processing millions of rows, reducing preprocessing time through optimized processing and caching.
• Developed an ML scoring system handling high-volume daily predictions with low latency, improving approval accuracy while maintaining model fairness.
• Created ensemble models for customer classification achieving strong performance metrics.
Machine Learning Engineer
Materials Center Leoben
Leading research company in Austria
Jun 2022 — Jul 2022Optimized and deployed edge AI models for Arduino Nano, reporting a 15% accuracy improvement and deployment to 100+ IoT devices.
• Optimized edge AI models (CNN, LSTM, VAE) for Arduino Nano, achieving a 15% accuracy improvement with 10ms inference latency and a 50KB memory footprint; deployed to 100+ IoT devices.
Data Science Developer
Bereke Bank
Top bank in Kazakhstan
May 2021 — Mar 2022Led a data analytics and sales-funnel initiative that reportedly generated $800,000 in product sales, reduced churn by 15%, and cut processing time by 20%.
• Led an end-to-end data analytics initiative processing 500GB across multiple sources, implementing a sales funnel strategy and feature engineering (3,000+ features) that delivered $800,000 in product sales while reducing customer churn by 15% and data processing time by 20%.
