Rajat Gosal is a Student Intern at the National Cybercrime Research & Innovation Centre in Dublin, Ireland. He is pursuing Advanced AI and Artificial Intelligence and Machine Learning at University College Dublin.

Mahesh Kanse is a seasoned professional with over 10 years of experience in application security, DevSecOps, and application management. Currently serving as Senior Vice President at Citibank in Dublin, Ireland, he combines his technical expertise with leadership to drive secure and efficient application development. Mahesh holds an educational background from Walchand College of Engineering, Sangli, and has also contributed to the Technological University of the Shannon. Beyond his corporate role, he is a technical product manager and tinkerer, actively engaged in projects involving large language models, agentic AI, Hugging Face, and Google ADK. A dedicated vegan, Mahesh blends innovation with a commitment to ethical living.

large language models, agentic AI, Hugging Face, Google ADK, application security, DevSecOps, application management

Llm, agentic ai, hugging face, google adk

Nethra, MSc Advanced AI student at UCD. Originally from Bangalore, currently in Dublin. I have spent the last few years building applied AI systems through internships, hackathons, and academic projects. Most of my work sits in the areas of LLMs, retrieval, and computer vision: things like guardrail pipelines for unsafe prompts, RAG-based ticket routing, multi-camera person re-identification, and on-device CNN inference for medical screening. Most of my time on a project goes into data work, evaluation, and the parts that fall apart pretty easily in the real world. Outside of building, I read a lot on infrastructure security and current affairs, and I am particularly interested in how AI is being used and misused in information ecosystems.

Looking to learn more about agentic systems, retrieval evaluation, and the operational side of LLM deployments: observability, drift detection, cost control. Also keen on misinformation and information integrity research, given my current side project. Happy to connect with people building in these areas, especially anyone working on RAG evaluation, agent frameworks, or LLMOps. Open to collaborators on side projects too if there is overlap in interests.

Just wrapped two MSc projects: a customer support ticket routing system using fine-tuned DistilBERT with a hybrid RAG pipeline (FAISS, BM25, reranker), and a multimodal breast cancer classification project, combining medical imaging with tabular data. Currently tinkering on a misinformation propagation fingerprinter, an early-stage system that traces how a claim mutates linguistically as it spreads across platforms. SBERT + FAISS for retrieval, Claude for drift analysis with propagation graphs.