Senior Applied Scientist

Microsoft Ads is seeking a Senior Applied Scientist with expertise in Large Language Model training and inference. In this role, you will be part of initiatives spanning pre-training, supervised fine-tuning, alignment using reinforcement learning and high-performance inference optimization. You will address challenges in multi-billion parameter model training, ensuring LLMs are robust, scalable, and optimized for production use. This is a unique opportunity to advance Microsoft Ads AI capabilities and solve real-world optimization problems for large-scale systems.

Microsoft’s mission is to empower every person and every organization on the planet to achieve more. As employees we come together with a growth mindset, innovate to empower others, and collaborate to realize our shared goals. Each day we build on our values of respect, integrity, and accountability to create a culture of inclusion where everyone can thrive at work and beyond.

Responsibilities

• LLM Training, Inference & Optimization.
• Train, and optimize large-scale transformer models for efficient training and inference.
• Distributed training (model/tensor/pipeline parallelism, sharding, data pipelines).
• Improve Training Efficiency.
• Create high-throughput, low-latency inference solutions.
• Partner with teams to integrate the model into Microsoft Ads scenarios.

Qualifications

Required Qualifications: 

• Bachelor's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 4+ years related experience (e.g., statistics predictive analytics, research)
  • OR Master's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 3+ years related experience (e.g., statistics, predictive analytics, research)
  • OR Doctorate in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 1+ year(s) related experience (e.g., statistics, predictive analytics, research)
  • OR equivalent experience.
• Experience training large-scale deep learning models (especially transformer-based architectures), with distributed computing (data parallelism, tensor parallelism, pipeline parallelism, model sharding, HPC techniques).
• Proficiency in PyTorch or TensorFlow for both training and inference.
• Reinforcement learning know-how for post-SFT alignment
• Solid understanding of model optimization for training and inference, including:
• Parameter-Efficient Fine-Tuning (PEFT) and Scaling techniques: LoRA, QLoRA, Adapters, MoE
• Quantization & Pruning: FP16/BF16 mixed precision, 4-bit & 8-bit quantization
• Memory-Efficient Techniques: ZeRO, FlashAttention, activation checkpointing
• Inference Optimizations: KV-cache optimization, speculative decoding, token streaming, distillation
• Experience with model compression & serving (e.g., ONNX, TensorRT, vLLM, FasterTransformer).
• Publication track record in top-tier AI/ML conferences (NeurIPS, ICML, ACL, ICLR, EMNLP)

Preferred Qualification:

• Experience in LLM pre-training using opensource frameworks like Megatron-LM
• Experience with inference acceleration: speculative decoding, FlashAttention, grouped-query/multi-query attention, dynamic batching.
• Practical experience with infrastructure orchestration for LLM inference: Triton Inference Server, ONNX Runtime, TensorRT, vLLM.

Microsoft is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to age, ancestry, color, family or medical care leave, gender identity or expression, genetic information, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran status, race, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable laws, regulations and ordinances. If you need assistance and/or a reasonable accommodation due to a disability during the application or the recruiting process, please send a request via the Accommodation request form.

Benefits/perks listed below may vary depending on the nature of your employment with Microsoft and the country where you work.

#MicrosoftAI #LLMTraining #EfficientInference #DeepLearning #RLAlignment #ModelOptimization