Skip to content
Carolina Code Conference
Served by Rust (axum) in 12ms

2026 speaker

Amit Kulkarni

Amit Kulkarni

Michelin NA Site Reliability Engineer

Tampa, FL

Amit is a Site Reliability Engineer specializing in enterprise observability and distributed systems reliability at Michelin North America. He designs and implements reliability frameworks for business-critical data pipelines, with a focus on bringing SRE principles to batch processing environments.

At Michelin, Amit pioneered the organization’s latency-based SLO governance framework, protecting over $2 billion in invoice revenue through proactive monitoring and automated incident detection. His work on data freshness as a reliability signal represents a novel approach to batch system observability, enabling end-to-end visibility across multi-stage enterprise workflows. His research on enterprise observability patterns has been productionized across multiple business domains at Michelin and shared with international teams. He holds expertise in Splunk, distributed systems monitoring, SLI/SLO frameworks, and reliability engineering methodologies.

Visit website

2026 · intermediate

Proactive Batch Observability: Automate File Freshness and SLA Monitoring

Modern observability focuses on APIs and microservices, yet many business-critical workflows still depend on batch jobs and file-based data pipelines. When these systems fail, they often do so silently only becoming visible after business impact.

This talk introduces a practical, tool-agnostic approach to batch observability by applying Site Reliability Engineering (SRE) principles. The core idea is simple but powerful:

In batch systems, data freshness is the equivalent of latency in real-time systems.

If data is late or missing, the system is effectively down from a business perspective.

Attendees will learn how to model batch workflows using event-driven observability, define expected vs. actual delivery SLAs, and measure reliability using a good event vs. total event SLI approach. These signals can then be translated into meaningful SLOs aligned with business outcomes.

The session also includes a brief real-world walkthrough demonstrating how these principles can be implemented to track end-to-end workflows, detect delays proactively, and provide a single, business-aligned view of reliability.

Why Attend

If your systems rely on batch jobs, scheduled workflows, or file transfers, this talk will help you:

  • Make batch systems observable and measurable
  • Detect issues before business impact occurs
  • Move beyond job-level monitoring to true reliability measurement
  • Apply SLI/SLO thinking to non-real-time systems

Key Takeaways

  • Model data freshness as a reliability signal (SLI)
  • Define and enforce SLAs based on data delivery
  • Connect technical signals to business impact
  • Apply a tool-agnostic framework across any observability stack

From the talk