research M.S. Thesis: PEFT and Jailbreak Robustness Does parameter-efficient fine-tuning make LLMs easier to jailbreak? SemEval 2026 Task 6 (CLARITY) Multi-seed DeBERTa ensembles for political response clarity and evasion classification SemEval 2026 Task 8 (MTRAGEval) Lightweight tri-fusion retrieval and faithful generation for multi-turn RAG UNLP 2026 Ukrainian document question answering with retrieval and reranking SemEval-2025 Task 10 Narrative framing in multilingual online news TurboQuant KV-cache compression for long-context language models work WorkWell Measure Studio Clinical quality measure platform built on FHIR and CQL RASS Retrieval-augmented semantic search over engineering issues JobOps Copilot Multi-tenant AI agent platform for job-search operations Patient Consent Smart Contracts Blockchain-based patient consent management