Jobs at Cerence Inc.

Build and own domain-specific AI agents for Cerence’s automotive assistant platform, from design through production deployment. Responsibilities include LLM prompting, function and tool calling, retrieval, conversational handling, cloud and edge compatibility, automated testing, behavioral evaluation, and production quality improvement. The role partners with product teams to translate user experience goals into reliable conversational agents.

Build and own domain-specific AI agents for Cerence’s automotive assistant platform. Responsibilities include designing LLM function-calling interfaces, integrating prompting, retrieval, tool selection, and conversational handling, and deploying shared agents across cloud and edge environments. The role emphasizes automated testing, behavioral evaluation, production reliability, collaboration with product teams, and agentic engineering practices.

Owns Cerence’s integrated global compliance management system across quality, HSE, security, privacy, legal/regulatory, and internal controls. Establishes policies, governance, risk prioritization, audit processes, and evidence frameworks; translates international regulations into operational controls; supports automotive cybersecurity, privacy, OEM, and SOX-related requirements; manages audits and corrective actions; and reports compliance risks to executives and the Audit Committee.

Design and implement advanced Generative AI models, optimize performance for LLMs, and collaborate with cross-functional teams to innovate AI solutions within the automotive sector.

Supports UX research, usability testing, participant recruitment, study logistics, findings analysis, and insight reporting. Assists with wireframes, user flows, mockups, prototypes, competitive analysis, and validation of conversational and multimodal automotive experiences. Collaborates with designers, researchers, product managers, and engineers while maintaining project artifacts in Figma, Confluence, and Jira.

Research Engineer role focused on helping Cerence advance voice and AI technologies for automotive applications. The position contributes to innovation supporting voice assistants used by global automakers and millions of vehicles. Specific daily responsibilities and first-year objectives are not provided in the description.

Own full sales and partnership cycle to bring voice AI to mobile service robots: engage senior robotics decision-makers, qualify and prioritize opportunities, close deals, and shape the commercial playbook while operating autonomously within a small technical team.

Own end-to-end software architecture for humanoid robot applications, design modular, production-ready platforms integrating robotics frameworks, AI, cloud/edge, and hardware constraints; lead technical decisions, guide POCs to deployments, and mentor a small engineering team.

Lead research and engineering work on LLM-based architectures for dialog, search, and destination entry. Optimize LLMs via prompting and fine-tuning, integrate them with ASR, TTS, and vision systems, and collaborate cross-functionally to deliver performant, resource-constrained solutions for automotive voice assistants.

Develop and maintain Python-based speech output tools, DevOps workflows, and CI/CD pipelines using Jenkins and GitLab. Analyze pipeline failures, improve tool performance, and support Docker and Kubernetes environments. Contribute to AI and MLOps initiatives involving speech and language model deployment, monitoring, and new product prototypes. The role also offers opportunities to work with Azure AI services, cloud-hosted machine learning, and generative AI tools.

Lead design and training of large-scale transformer and hybrid foundation models across text and multimodal domains. Own architecture choices, diagnose large-scale training instabilities, optimize scaling tradeoffs (data, compute, model), design loss and alignment strategies, and execute distributed training (FSDP/ZeRO, tensor/pipeline parallelism) with mixed precision to build in-house foundation model expertise.

Integrate and deliver voice recognition software for automotive, collaborating with clients and internal teams, while setting up CI/CD environments and handling testing.

Lead development and optimization of high-performance LLM inference pipelines across data center, edge, and embedded platforms. Improve latency and throughput via custom CUDA kernels, quantization, KV-cache and batching strategies, and deploy efficient, production-ready runtimes without external vendor dependency.

Lead technical engagements with robotics customers to design and integrate voice AI into mobile robots, own POCs and technical evaluations, clear deployment roadblocks, collaborate with sales/product/services, and feed customer learnings into the product. Travel across Europe for onsite evaluations (20–30%).

Customer-facing sales role responsible for owning direct opportunities and managing channel/distributor relationships to drive revenue. Duties include discovery, proposals, demos, negotiation, pipeline management, partner enablement, co-selling, event representation, and cross-functional coordination to support go-to-market and product launches. Targets include bookings, distributor-sourced pipeline growth, and executing partner workshops.

Develop speech dialog system applications for Android, implement VUI/GUI, ensure app performance, collaborate on features, fix bugs, and maintain code quality.

The Principal ASR Engineer will develop and enhance ASR systems with a focus on Japanese language, utilizing data for training and testing while collaborating with teams to optimize performance.

The Sr. Principal TTS Researcher will develop TTS systems, utilizing expertise in frontend/backend components, machine learning, and NLP while leading cross-functional teams.

Design and implement advanced generative AI models, optimize deployment in production, collaborate with teams, and stay updated on AI advancements.

Design, build, and operate distributed GPU training infrastructure for large neural networks. Optimize multi-node/multi-GPU execution, diagnose compute/memory/network bottlenecks, improve stability and fault tolerance, and partner with research teams to productionize large-model training pipelines for reliable, high-throughput training at scale.