This role focuses on applying advanced machine learning and audio processing techniques to solve complex, real-world problems with meaningful user impact. If you enjoy working with speech, text, and multimodal data and thrive in a fast-paced, highly collaborative environment, we’d love to connect. Key Responsibilities • Design and develop a conversational AI assistant powered by large language models (LLMs) for embedded systems. • Benchmark and evaluate voice-native models (e.g., Moshi, SesameAI) for performance in automotive contexts. • Fine-tune foundational models for speech-based tasks with domain-specific adaptation. • Build API layers for controlling vehicle systems via natural language interfaces. • Implement performance optimizations tailored to the Qualcomm SA8255P hardware platform. • Maintain and enhance Python-based frameworks for audio preprocessing, model orchestration, and deployment on edge devices. • Conduct architecture analysis, model benchmarking, and performance tuning. • Manage the full ML lifecycle: data prep, feature engineering, model experimentation, and production readiness. • Contribute to Agile teams by prototyping, building, and iterating on models that translate into scalable solutions. Minimum Qualifications • 3+ years of professional experience in machine learning and AI product development. • Strong expertise in voice-to-voice systems and speech-focused machine learning architectures. • Experience with speech recognition (ASR) and text-to-speech (TTS) systems using Transformer models. • Proficient in model optimization for edge deployments, including quantization and hardware-aware tuning. • Solid grounding in digital signal processing and acoustic modeling for audio inputs. • Practical knowledge of PyTorch or TensorFlow and related audio libraries (e.g., TorchAudio, Librosa). • Familiarity with Huggingface Transformers and distributed training pipelines. • Understanding of multimodal learning and fusion techniques (e.g., cross-modal attention). • Experience designing and implementing training pipelines for large-scale model experimentation. • Comfortable working with complexity, performance constraints, and ambiguous problem sets. • Proven problem solver with a bias for execution and experimentation. • Effective communicator and collaborator across product, engineering, and research functions. • Ability to work independently while driving results within a cross-functional team.