Michael Kim

I’m a Senior Software Engineer with 12+ years of experience building and scaling AI/ML systems at Datadog, Salesforce, and Amazon. I specialize in end-to-end ML platform engineering, from real-time inference and LLM observability to MLOps pipelines and multi-tenant AIOps.

I’m known for translating cutting-edge AI capabilities into reliable, low-latency production infrastructure that directly reduces operational burden for engineering and SRE teams. Across enterprise-scale, cloud-native environments, I’ve delivered systems that process trillions of data points daily.

At Datadog, I led backend architecture for AIOps and LLM Observability platforms that power real-time anomaly detection, AI-driven root cause analysis, and generative AI monitoring across thousands of customers. I architected AIOps backend processing trillions of points daily for Watchdog (sub-second latency), engineered event-driven pipelines for Bits AI (reducing customer MTTR by 20% via LLM-driven remediation), and built telemetry to monitor token usage and cost for OpenAI and Anthropic.

Before that, at Salesforce, I built the core MLOps and multi-tenant infrastructure behind Einstein AI, training over 900,000 customer-specific ML models per hour. I also engineered scalable workflows for the ML Lake on AWS S3 and Apache Iceberg, designed distributed scheduling with AWS Lambda to orchestrate hundreds of thousands of parallel training jobs, and optimized real-time prediction serving infrastructure for low-latency outcomes inside Salesforce applications.