Agamyrat Akmyradov

Led the design and deployment of an agentic, multimodal healthcare AI platform for real-time clinical decision support and workflow automation. Built end-to-end data pipelines from EHR (FHIR) and document inputs, developed hybrid RAG and multi-agent orchestration, and integrated MCP to connect LLMs with clinical data sources. Implemented fine-tuning and optimized inference, plus evaluation and monitoring with MLOps guardrails for GDPR/HIPAA.

Developed NLP-driven automation for financial documents, including classification, summarization, and anomaly detection. Built conversational and retrieval-based assistant prototypes to automate a large share of customer support interactions. Implemented scalable data and MLOps workflows using PySpark plus Kubeflow, Terraform, and CI/CD, with human-in-the-loop evaluation and compliance-oriented model alignment.

Built end-to-end ML solutions for fraud detection, churn prediction, and credit scoring using classical models and ML tooling. Created ETL and transformation pipelines for large transaction volumes, and deployed time-series forecasting models to improve KPI predictions. Designed analytics data warehouses and set up CI/CD for reproducible ML training and continuous delivery.

Developed and maintained backend APIs in Python using Flask with PostgreSQL to support analytics platforms. Built interactive dashboards using JavaScript, D3.js, and Plotly to improve reporting efficiency. Implemented data processing and automation for client needs and created ML utilities using scikit-learn and Spark for enhanced analytics and predictive features.