- 🔬 Researcher at Heinz Nixdorf Institute/DICE Group, Paderborn University
- 🎓 Master's in Computer Science from RPTU Kaiserslautern – Landau
- 🧠 Passionate about building human-aware AI systems that are explainable, adaptive, and trustworthy
- 💡 Researching Retrieval-Augmented Generation (RAG) evaluation methodologies, comparing traditional metrics with LLM-as-judge approaches
- 🛠️ Experienced in Deep Learning, LLM Evaluation, Emotion Recognition, and Meta-Learning
- 🏢 Previously contributed to real-world AI at Capgemini, Fraunhofer IEE, and research at DFKI
- ☕ Fueled by curiosity, creativity, and coffee
- Languages & Frameworks: Python | Java | C++ | Flask | PyTorch | TensorFlow | Keras | Streamlit
- Machine Learning & LLMs: Deep Learning | Computer Vision | Affective Computing | LLMs | AutoGen | Groq | LangChain | RAG
- Cloud & DevOps: AWS (Redshift, Glue, Lambda, S3, SQS, SNS) | Docker | Grafana | Loki | Github Actions | KubeRay | MiniKube
- Databases & Visualization: PostgreSQL | MySQL | TimescaleDB | Tableau | Power BI | AWS Quicksight
- Human-in-the-Loop Annotation for Image-Based Engagement Estimation - HCI International 2025 (Gothenburg, Sweden)
- Stress Detection using Deep Learning and IOT - International Journal of Research in Engineering, Science and Management (Vol. 2, Issue 8)
- English Accent Classifier : A Streamlit-based tool that analyzes spoken English from public videos to classify accents like American, British, or Indian using Whisper and SpeechBrain. Outputs include a confidence score and visual summary, ideal for screening or language evaluation.
- Stylized RAG: AI-Powered Text Style Transformer : A stylized Retrieval-Augmented Generation (RAG) application built using Streamlit, LangChain, and Wikipedia, designed to deliver aesthetically engaging responses. It supports dynamic retrieval pipelines and integrates creative response formatting for enhanced readability and user interaction.
- Multi-Agent Meal Planning System : An AI-powered system that uses specialized agents to generate meal plans based on diet, budget, and calorie goals. Built with Flask, it supports multiple diets and creates a consolidated shopping list from the selected meals.
- ai-invoice-processing : An AI-supported tool that fetches invoice emails, extracts key details using Google Vision API, and stores them in MongoDB. Includes a Flask dashboard to review, edit, and track invoices with overdue or recurring status.
