John Aghaei

Integrated Claude Code and AI coding assistants to accelerate implementation, debugging, automated testing, and code refactoring workflows. Designed and deployed agentic and RAG-based AI assistants for enterprise clients with codebase-aware retrieval, prompt orchestration, and tool calling from architecture to production. Built AI evaluation and benchmarking pipelines with RAGAS and Ranx to measure agent quality and reliability for end-to-end pipeline success. Analyzed requirements, selected models/tools, delivered LLM-powered client features, fine-tuned small domain models, and optimized model selection to reduce inference cost while meeting performance needs. Collaborated with client engineers on requirements, prototypes, and iterative improvements.

Founded and scaled an online IT education platform to 300,000+ students across 30+ programming and software development courses. Produced 240+ hours of technical video courses on YouTube, contributing 1M+ watch hours, 8M+ views, and 100,000+ subscribers, demonstrating strong technical communication through teaching and content creation.

Delivered 500+ hours of IT training in programming and software development for students and professionals, supporting technical upskilling and applied learning through hands-on instruction.