Member of Technical Staff - Multimodal - Mai Superintelligence Team

Microsoft Microsoft · Big Tech · Mountain View, CA +4 · Software Engineering

This role is focused on building and advancing large-scale foundation models, with a specific emphasis on multimodal capabilities and ensuring AI systems are controllable, safety-aligned, and anchored to human values. The position involves algorithm development, model architecture design, experimentation, data pipeline innovation, and improving training/deployment efficiency, aiming to push the frontier of AI responsibly.

What you'd actually do

  1. Develop algorithms, design model architectures, conduct experiments, champion measurement and evaluation, innovate datasets and data pipelines.
  2. Improve training and deployment efficiency, paying careful attention to detail, persevering, and learning from everyone’s attempts whether successful or not.
  3. Follow a rigorous data-driven approach grounded in meticulous ablation studies and scientific analysis.
  4. Innovate and iterate over ideas, prototypes, and product.
  5. Collaborate closely with teams on infrastructure, data engineering, pre-training, post-training, and product feedback.

Skills

Required

  • Bachelor's Degree in AI, Computer Science, Data Science, Statistics, Physics, Engineering, or related technical discipline
  • 4+ years technical engineering experience with coding in languages including, but not limited to, C, C++, C#, Java, JavaScript, or Python

Nice to have

  • Master's Degree in in AI, Computer Science, Data Science, Statistics, Physics, Engineering, or related technical discipline
  • 8+ years technical engineering experience with coding in languages including, but not limited to, Python and common data libraries (Pandas, NumPy, etc.)
  • Experience with large-scale AI systems — design and deployment of distributed architectures, multimodal or conversational models; proficiency with ML frameworks (e.g., PyTorch, TensorFlow) and cloud/HPC environments (e.g., Azure).
  • Expertise in data engineering for foundation models — multimodal dataset design, curation, annotation pipelines, quality evaluation, bias detection, and understanding of privacy, compliance, and Responsible AI principles.
  • Background in LLM interaction and deployment — practical work in prompt engineering, safety-aligned evaluation, and integration of conversational AI into production systems.
  • Cross-functional collaboration and communication — ability to produce clear technical documentation, partner with engineering, product, and design teams, and contribute to knowledge sharing; demonstrated application of emerging AI technologies and best practices.

What the JD emphasized

  • proven expertise, demonstrated through impactful publications or technical leadership on high-scale projects
  • large-scale AI systems
  • multimodal or conversational models
  • data engineering for foundation models
  • LLM interaction and deployment

Other signals

  • building foundation models
  • pushing boundaries of scale, performance, and deployment
  • frontier AI systems
  • multimodal
  • humanist superintelligence
  • controllable, safety-aligned, and anchored to human values