At FlyRankAI, I analyze content and performance datasets to find actionable signals for SEO opportunity ranking. I build transparent rule-based machine learning baselines with scoring, reason codes, and ranked action queues.
I audit signals and validate hypotheses with statistical summaries and bucket analysis. I also apply leakage-aware feature selection and honest validation practices to machine learning workflows.
On my Multimodal AI Assistant project, I built document- and image-based question answering with Python, FastAPI, Streamlit, Gemini, and RAG. I added FAISS-based vector retrieval and connected separate document and image processing workflows through APIs.
I also added authentication, chat history, database integration, and PDF report generation to the assistant. For my Movie Recommender System project, I developed preprocessing and similarity-based recommendation logic to retrieve films from user preferences or a selected movie.
