Level AI is a Mountain View-based conversational intelligence company building the AI platform that powers modern contact centers. Our systems analyze millions of customer interactions to surface insights, automate quality assurance, and drive real-time agent guidance. We work at the frontier of applied speech and language AI, and our research directly shapes products used by enterprise customers every day.
We are looking for a Principal AI Research Engineer to set the technical direction for our most ambitious AI initiatives. You will lead research on large language models, speech understanding, and retrieval systems—then translate that research into production capabilities at scale. This is a senior individual-contributor role for someone who thrives at the intersection of cutting-edge research and real-world deployment, and who can raise the technical bar across the entire AI organization.
What You'll Do
Define and drive the research roadmap for core AI capabilities spanning LLMs, ASR/speech understanding, retrieval-augmented generation, and agentic systems.
Design, train, fine-tune, and evaluate state-of-the-art models, taking them from prototype to robust, production-grade systems.
Lead the architecture of scalable training and inference pipelines that operate reliably on real customer-scale data.
Establish rigorous evaluation frameworks, benchmarks, and quality bars for model performance, safety, and reliability.
Partner closely with product, platform, and ML engineering teams to ship research into customer-facing features.
Mentor senior and mid-level engineers, set technical standards, and act as a force multiplier across the AI org.
Stay ahead of the field, evaluate emerging techniques, and represent Level AI's technical work externally where appropriate.
What You'll Bring
Degree in Computer Science, Machine Learning, or a related field, or equivalent research experience; typically 7+ years applying ML/AI in production settings.
Deep expertise in modern deep learning, with strong command of LLMs, transformers, and at least one of speech/ASR, NLP, information retrieval, or multimodal modeling.
A track record of taking research from idea to deployed, scaled product—not just publications, but shipped impact.
Expert-level proficiency in Python and a major deep learning framework (PyTorch preferred), with strong software engineering fundamentals.
Hands-on experience training and serving large models, including distributed training, optimization, and efficient inference.
Demonstrated technical leadership: setting direction, mentoring engineers, and influencing across teams.
Clear communicator able to convey complex technical ideas to both research and product audiences.
Nice to Have
Experience in conversational AI, contact center, or speech analytics domains.
Publications at top-tier venues (NeurIPS, ICML, ACL, EMNLP, Interspeech) or significant open-source contributions.
Experience with agentic frameworks, tool use, and orchestration of LLM-based systems.
Familiarity with model safety, alignment, and responsible AI evaluation.
Why Level AI
Work on AI problems that ship to real enterprise customers at scale, with measurable impact.
Join a small, high-caliber team where your technical decisions shape the product and the org.
Competitive compensation, meaningful equity, and comprehensive benefits.
A culture that values research depth and pragmatic, production-focused execution in equal measure.
Level AI is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.
Lead the technical direction of production machine learning systems for deposition and meeting products, building grounded LLM pipelines and evaluations while guiding architecture, reliability, and engineering practices.
Own cross-team machine learning architecture and evaluation standards, guide agentic AI design, and strengthen production practices across AI teams while ensuring safe, compliant deployment.
Lead the technical strategy and end-to-end development of foundation models, from data curation and pretraining through production serving, evaluation, and monitoring, while guiding teams across Plaid.
Leads the strategy, architecture, and deployment of secure, scalable machine learning systems, advancing AI safety evaluations and multi-agent tooling while advising clients and mentoring engineering teams.
Lead end-to-end development of production machine learning systems for cash-flow underwriting, shaping technical direction and turning transaction data into fair, accurate credit decisions and measurable member outcomes.
