Bharath Raj

I’m an AI Engineer with hands-on internship experience in ML systems, fine-tuning, and LLM-based pipelines. I’ve built production-quality workflows with Python and the Hugging Face/transformer ecosystem, with a strong focus on evaluation pipelines, lean inference, and measurable iteration.

In my ML internships, I fine-tuned and optimized classification models using scikit-learn hyperparameter tuning, improving F1-score by 12% and reducing prediction latency by 18%. I also streamlined models by identifying 4 key predictors (simplifying inference with no accuracy loss) and improved runtime performance by resolving codebase bottlenecks within tight deadlines. On projects, I’ve delivered end-to-end results like an XGBoost churn pipeline (89% accuracy with SHAP-based interpretability) that reduced quarterly churn by 12%, and a multi-agent LLM automation workflow using Ollama, REST APIs, and n8n that cut manual content effort by 80%.