Engineering Manager - AI Engineering

About the role

Meesho is expanding its AI infrastructure team to support high-scale e-commerce operations. You will manage a technical group focused on building and optimizing the platforms that power real-time deep learning inferences for millions of users.

Key facts

What you'll do

• Direct the technical strategy and execution for a team of AI engineers.
• Design and scale infrastructure for multi-region model inference, distributed training, and feature engineering.
• Improve inference performance through GPU kernel tuning, quantization, and memory optimization.
• Refine open-weight models using techniques like speculative decoding, distillation, and serving-engine adjustments.
• Develop automated, agent-driven workflows to increase productivity in model training and rollout.
• Collaborate with Data Science and Product teams to deploy AI solutions that impact the consumer experience.
• Manage team operations including hiring, performance reviews, and goal setting.

Requirements

• Bachelor or Master degree in Computer Science or a related discipline.
• 9+ years of professional software engineering experience.
• 2+ years of experience in a management role.
• Proficiency with LLM inference tools such as TensorRT-LLM, vLLM, and SGLang.
• Experience with low-latency model serving in production environments.
• Knowledge of distributed training frameworks like PyTorch FSDP, DeepSpeed, Megatron, or Ray.
• Experience managing GPU fleets via Kubernetes, including scheduling and multi-region deployment.
• Familiarity with big-data and streaming technologies like Spark or Flink.
• Strong coding skills in Python and systems-level languages such as C++, Go, or Rust.

Nice to have

• Public contributions to open-source ML infrastructure or inference engines.
• Experience managing GPU costs and efficiency on cloud platforms.
• Background in building platforms for large-scale consumer applications.
• Expertise in ML observability, reliability, and incident management.

Skills & tools

• Python, C++, Go, Rust
• TensorRT-LLM, vLLM, SGLang
• PyTorch FSDP, DeepSpeed, Megatron, Ray
• Kubernetes, GKE
• Spark, Flink
• CUDA, GPU programming