At Micro1, I review AI-generated text and multimedia outputs against safety and policy guidelines, maintaining 99%+ accuracy. I also categorize complex content and flag inconsistencies, errors, and policy violations.
At IQBanana, I integrated LLM tools to extract and categorize insights from unstructured data. I managed ETL pipelines and logged data quality issues in Jira.
On MindSphere, I trained a CatBoost model on 280K+ records, achieving 92.4% accuracy, and used SHAP for feature attribution. I also automated evaluation reports with Flask and SQLite.
I developed Insightly AI to generate natural-language summaries and KPI recommendations from datasets, and evaluated its Groq LLM outputs for accuracy and policy compliance. I also co-authored a research publication on machine learning-driven mental wellness assessment.
