• Scaling GPU Workloads on AKS: Fair Sharing, Preemption, and MIG

  How to run multi-tenant GPU workloads on Kubernetes with Kueue for fair scheduling and NVIDIA MIG for GPU slicing.

• Dec 2023 - Current
  Microsoft
  Applied AI, Technical Lead

  Working on applied AI — taking models from “cool demo” to production. Building with LLMs, Azure AI, and whatever open source tools fit the problem. Less PowerPoint, more code.

• Sep 2022 - Nov 2023
  Amazon Web Services (AWS)
  Solutions Architect

  Helped teams figure out how to build things on AWS without over-engineering them. Lots of LLM workloads, microservices, and the occasional “please don’t put everything in one Lambda.”

• Visionary Lab

  Create high-quality visual content with GPT-Image-1 and Sora on Azure OpenAI — tailored for professional use cases.
• Synthetic Data Generation (Microsoft Foundry)

  A feature in Microsoft Foundry for generating synthetic datasets to augment training data and validate ML models.
• Insurance Multi-Agent Claims

  An agentic AI demo that uses multi-agent collaboration to streamline end-to-end insurance claims processing.

Let's Connect

• github /
• linkedin /