Nitin Saksena

Title of the Talk :
Agents Beyond LLMs: A Roadmap from Pilot to Production of Agentic AI Platforms

Abstract of Talk:

A practical, end-to-end playbook for moving agentic AI from exciting pilots to reliable production—using the right mix of LLM agents and beyond-LLM methods, with governance that prevents “agent anarchy.” Agent platforms are booming but fragmented; “agent-washing” is common; and many deployments stall due to reliability, integration risk, and lack of governance. This talk helps attendees make clear platform choices and scale responsibly.

Key takeaways :

  • A decision framework for choosing the right agent approach (LLM-only vs. hybrid vs. beyond-LLM) for a given workflow.
  • A pilot-to-production roadmap with stage gates, metrics, and anti-patterns to avoid.
  • A set of production patterns for reliability, evaluation, and cost control.
  • A lightweight governance model that enables scale while reducing security/compliance risk.

Bio-data:

Nitin Saksena is a Senior Director of Engineering at Albertsons Companies, where he leads enterprise-scale architecture and engineering initiatives across eCommerce, digital shopping experiences, marketing, retail media, merchandising, supply chain, and pharmacy platforms. With over 20 years of experience in large-scale retail and digital transformation, Nitin has been instrumental in architecting mission‑critical systems. A recognized expert in event-driven architecture, data streaming, real-time analytics, and AI/ML-powered personalization, Nitin focuses on building resilient, high‑throughput platforms that translate data into measurable business and customer outcomes. His work bridges enterprise architecture with applied machine learning to enable intelligent, omnichannel retail experiencesBeyond his industry leadership, Nitin is a Senior Member of IEEE, brings a practitioner’s perspective combining strategy, architecture, and hands‑on execution to share actionable insights on scaling AI, modern data platforms, and real‑time systems in complex enterprise environments