Shahnawaz Alam

About Shahnawaz Alam

I build software infrastructure for AI systems - inference serving for large language models with agents orchestration, GPU-accelerated processing pipelines, and computer vision at scale.My current work focuses on LLM inference infrastructure: serving runtimes for 100B+ parameter models, designing platform strategies for multi-node inference with sub-minute latency constraints for agent execution, and benchmarking systems under real production conditions. I also build event-driven ML pipelines on AWS - GPU-backed video processing, model monitoring, and automated retraining systems deployed across multiple regions.Before Amazon, I built real-time computer vision systems - multi-camera tracking on edge devices, scene understanding for autonomous driving, and perception algorithms for robotic manipulation. Earlier in my career, I spent five years in research on biomedical signal classification and mobile sensing, publishing at KDD, ICASSP, and CinC, and was granted four US patents.MS in Media Informatics, RWTH Aachen University. 10+ peer-reviewed publications.

Experience

1. Sde, Ai Infrastructure

   Amazon

   Jan 2026 — Present · Bellevue, WA, US

   Building AI infrastructure spanning LLM inference for Agentic systems and production ML pipelines.Designed the inference scaling strategy for a multi-step agentic LLM platform.

Education

• RWTH Aachen University

  Master of Science - MS, Media Informatics

Skills

• Mac Os
• Matlab
• Windows Phone
• Data Structures
• Android Development
• C
• Java
• Ndk
• Unity3d
• C#
• C++
• Javascript
• Ubuntu
• Eclipse