AI Research Scientist - SLM and Voice Advanced (Senior)

Advanced Model Development: Design and implement novel machine learning models with a special focus on Small Language Models (SLMs) and Automatic Speech Recognition (ASR). Develop robust speech processing for noisy environments (ASR, TTS) and domain-specific language models. Model Optimization and Alignment: Lead efforts in model compression and efficiency. Utilize advanced techniques such as knowledge distillation, quantization, and pruning to create lightweight models for resource-constrained edge devices. Implement and experiment with state-of-the-art model alignment techniques, including RLHF and DPO. Research and Innovation: Push the boundaries of what is possible in AI by translating academic theory into production-grade technology. Conduct research that advances the state-of-the-art in SLMs and ASR, with opportunities to patent inventions and publish findings at top-tier conferences. Data-Centric AI: Architect and manage data pipelines for large-scale datasets. Implement algorithms for automatic data curation, cleaning, and pre-processing to ensure high-quality inputs for model training and fine-tuning. Prototype to Production: Lead the transition of research prototypes into scalable, production-ready solutions by collaborating closely with engineering and product teams. Collaboration and Mentorship: Collaborate effectively in a team environment on shared codebases using Git/GitHub. Mentor junior scientists and engineers, fostering a culture of technical excellence, curiosity, and innovation. Education: PhD or Master\u0027s degree in Computer Science, Artificial Intelligence, a related technical field, or equivalent practical experience. Proven experience in designing and implementing machine learning models, particularly in Small Language Models (SLMs) and/or Speech Recognition (ASR). Demonstrated, hands-on experience with advanced model optimization techniques, including knowledge distillation, quantization, and pruning. Deep understanding and practical experience with model alignment methods such as RLHF, DPO, or similar techniques. Programming: Strong proficiency in Python and extensive experience with ML frameworks such as PyTorch, TensorFlow, or JAX. Fundamentals: Solid theoretical and practical understanding of deep learning, reinforcement learning, and statistical modeling. A strong track record of publications in top-tier AI conferences (e.g., NeurIPS, ICML, ICLR, ACL, Interspeech, CVPR). Familiarity with the Hugging Face ecosystem (Transformers, Datasets, TRL). Experience using tools such as Hydra for configuration management and Weights \u0026 Biases for experiment tracking. Familiarity with MLOps best practices (Docker, Kubernetes, W\u0026B) to support reproducible research and scalable deployment. Strong communication skills to articulate complex technical concepts to product stakeholders.