• Design, develop, and refine large language model (LLM) workflows, including context engineering, prompt design, and evaluation frameworks to steer and improve model behaviors. • Build language processing components for features such as intent detection, entity recognition, summarization, retrieval-augmented generation (RAG), and conversational response quality. • Develop speech-to-text (ASR) and text-to-speech (TTS) workflows and evaluation frameworks, bridging audio-feature/signal-level processing with LLM-driven reasoning and orchestration. • Fine-tune and evaluate models using quantitative and qualitative metrics to ensure robust performance across tasks. • Analyze model outputs and conversational data to identify patterns, gaps, and failure modes, translating findings into actionable improvements. • Define and apply linguistic evaluation criteria to ensure tone, clarity, intent understanding, and contextual accuracy. • Experiment with prompt structures, retrieval strategies, and linguistic patterns to improve accuracy and robustness. • Drive R&D-style exploration on cutting-edge speech and language systems where best practices are still emerging, rapidly prototyping novel approaches and validating them through rigorous experimentation. • Lead data preprocessing, annotation, and language dataset creation, building reliable training and evaluation corpora. • Design experiments to test model adaptations and new techniques, tracking performance and iterating based on data insights. • Collaborate with software developers to integrate language models into production systems and ensure scalable deployment. • Build tooling for model evaluation, monitoring, and continuous improvement pipelines. • Support performance optimization, model serving architecture, and infrastructure integration. • Partner with product managers, conversation designers, UX researchers, and stakeholders to connect language capabilities with business objectives. • Serve as the NLP & language subject-matter expert within multidisciplinary teams. • Document methodologies, evaluation findings, best practices, and language guidelines to promote shared knowledge and reproducible workflows. • Present results and recommendations clearly to internal and external stakeholders.