At LegitAI Solutions, I build Python applications that connect LLM APIs with structured data pipelines, schema validation, and automated production workflows.
I engineered a multi-source pipeline that transformed 3,735 raw records from six sources into 1,402 cleaned, validated, and deduplicated records.
I've also tested REST APIs at Sabrhub using Postman, Newman, and GitHub Actions CI, while building automated validation suites and documenting edge cases. Earlier, I developed Python prototypes for data masking, vulnerability scanning, encrypted activity logging, and multi-user authentication.
My independent research projects focus on rigorous AI evaluation: I built monitoring and benchmarking frameworks using deterministic grading, McNemar tests, bootstrap confidence intervals, FastAPI, Docker, and automated tests.
