Gagandeep Shivanna

Owned end-to-end handling of assigned tickets, from creating structured training data to performing QA and reasoning-based evaluation, ensuring each output met accuracy, relevance, and safety standards. Applied domain expertise and critical reasoning to compare and rate model responses in RLHF-style side-by-side evaluation, selecting and annotating the stronger output to reinforce desired model behavior. Maintained rigorous quality benchmarks through gold-standard tasks and self-audit of borderline cas

Built and evaluated NLP pipelines for document classification and information extraction under Prof. Dr. Alexander Iliev. Applied prompt engineering and retrieval-augmented generation (RAG) to improve extraction accuracy on domain documents. Synthesized research across cognitive science and AI ethics into structured academic writing

• Collaborated with the Java backend development teams such as ”Zaxby’s”, ”Tonic” projects. • Assisted senior developers in implementing new features and modifying existing functionality to meet project requirements. • Contributed to the design and development of backend services, ensuring high performance and scalability. • Took charge of designing, debugging, developing, and testing APIs, ensuring their robustness, reliability, adherence to project requirements, interacting with databases, guarantee

•Hands-on experience with AWS, Azure, and Google Cloud Platform •Completed multiple cloud certifications •Developed and led AI-driven projects and research papers Notable Projects: •Detection of Deepfakes in Art and Media - Trained ML models to detect manipulated content using CNN and GAN arch