Hey there, I amΒ Pooja
π M.Sc. Data Science graduate | Berlin, Germany
π€ Focused on Machine Learning, Generative AI, LLMs, RAG, and end-to-end AI applications.
π» Nearly 4 years of professional experience as a QA Automation Engineer at Finastra(LOAN IQ), a fintech company.
π³ Thesis: designed & evaluated 60 ML pipelines for credit card fraud detection on imbalanced data (best PR-AUC: 0.81)
π Building hands-on projects using Python, FastAPI, Docker, scikit-learn, FAISS, and CI pipelines.
π± Actively contributing to open-source projects and strengthening my AI engineering skills.
π Based in Berlin, open to opportunities across Germany
π« Reach me: Email:mbpoojaraju012@gmail.com | LinkedIn:Pooja Medigeshi Basavaraju
Tech Stack: Python, Gemini API, FAISS, FastAPI, Streamlit, Docker, GitHub Actions
- Built an end-to-end Retrieval-Augmented Generation (RAG) application for question answering over technical PDF documents.
- Implemented document extraction, chunking, embeddings, vector search, and LLM-based answer generation with source references.
- Developed REST API endpoints using FastAPI and an interactive Streamlit interface.
- Containerized the application with Docker and implemented automated CI checks using GitHub Actions.
Tech Stack: Python, scikit-learn, XGBoost, Pandas, SMOTE, FastAPI
- Designed and evaluated 60 ML pipeline configurations for highly imbalanced credit card fraud detection.
- Applied feature engineering, chronological train-test splitting, and multiple imbalance-handling techniques.
- Achieved a best PR-AUC of 0.8107, precision of 0.8871, and recall of 0.7432.
- Developed a FastAPI inference application for serving fraud predictions.
Tech Stack: Python, Pandas, Matplotlib, Log Parsing
- Developing a production incident diagnosis system using real-world OpenStack logs.
- Implemented log ingestion, parsing, structured preprocessing, and exploratory analysis.
- Analyzed log severity distributions and component-level patterns to support incident investigation.
Status: In Progress
Tech Stack: Python, FastAPI, Pytest, Git, GitHub Actions
- Contributed to a transaction-policy validation project.
- Implemented JSON serialization for transaction decisions.
- Developed FastAPI endpoints and unit tests for transaction-policy checks.
- Practiced Git branching, pull requests, code review, and automated testing.
