> Markdown version of [/videos/100369-acquired-live-interview-justin-boitano-vp-enterprise-ai-products-nvidia?t=1604](https://www.wearedevelopers.com/videos/100369-acquired-live-interview-justin-boitano-vp-enterprise-ai-products-nvidia?t=1604). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Acquired Live Interview: Justin Boitano (VP, Enterprise AI Products, NVIDIA)

Justin Boitano reveals how fine-tuned open-source AI now outperforms proprietary frontier models. Discover how NVIDIA's kernel-level infrastructure enables organizations to safely orchestrate millions of autonomous enterprise agents.

- **Speakers:** [Justin Boitano](https://www.wearedevelopers.com/@justin-boitano), [Ben Gilbert](https://www.wearedevelopers.com/@ben-gilbert), [David Rosenthal](https://www.wearedevelopers.com/@david-rosenthal)
- **Event:** World Congress 2026 North America
- **Published:** September 24, 2026
- **Duration:** 27:52
- **URL:** https://www.wearedevelopers.com/videos/100369-acquired-live-interview-justin-boitano-vp-enterprise-ai-products-nvidia

## Summary

The enterprise transition from experimental AI to scalable, agent-driven architectures requires bridging probabilistic models with deterministic governance. NVIDIA's VP of Enterprise AI Products, Justin Boitano, outlines how autonomous agents are reshaping software development and business operations, viewing AI safety primarily as an engineering challenge. To address this, NVIDIA introduced OpenShell, an open-source framework that enforces granular access controls and formal verification at the Linux kernel level. This infrastructure layer ensures that agents interacting with sensitive business systems remain strictly within their scoped permissions.

Open-source AI models are rapidly closing the performance gap with proprietary frontier models, now trailing by only a few months. By post-training open foundational models like Nemotron on domain-specific enterprise data, organizations can achieve better-than-frontier capabilities at a fraction of the cost. Cybersecurity firms are leveraging this approach to build specialized models for red teaming and infrastructure hardening. Similarly, highly regulated industries constrained by data residency and compliance rules are utilizing confidential computing partnerships and localized open-source deployments. This allows them to bring powerful AI capabilities directly into secure data centers without risking external data exposure.

Looking forward, the architecture of enterprise AI will demand extreme elasticity, seamlessly routing workloads from local developer workstations, such as DGX Spark, to on-premise AI factories and cloud environments based on cost and performance needs. The future workplace will be dominated by agent-to-agent communication, with massive enterprises eventually orchestrating millions of internal agents to automate complex workflows. Crucially, silicon providers must continuously optimize their software stacks and build specialized hardware kernels to maximize performance-per-watt for every new major model release. Ultimately, collaborative transparency through initiatives like the Open Secure AI Alliance will be vital for managing agent safety and infrastructure integrity at scale.

**Keywords:** enterprise AI agents, deterministic access controls, AI security engineering, openshell framework, linux kernel verification, domain-specific model post-training, cybersecurity red teaming, nemotron optimization, data residency compliance, confidential computing infrastructure, agent-to-agent communication, inference compute routing, hardware performance-per-watt, AI infrastructure elasticity, DGX workstations, open secure AI alliance

## Chapters

1. **Evolving from a graphics company to accelerated computing leader** (00:00) — Returning to a rapidly scaling organization driven by a founder's long-term infrastructure vision.
1. **Transitioning the industry toward an era of automated agents** (02:29) — The technology stack is evolving to support automated agents that allow even young developers to build and collaborate.
1. **Securing artificial intelligence agents with kernel level deterministic controls** (04:27) — Enforcing network policies and file system access ensures automated systems only execute scoped tasks within regulated environments.
1. **Leveraging open foundation models for enterprise cybersecurity applications** (07:32) — Specialized organizations can post-train open models on domain specific data to outperform general frontier models at a fraction of the cost.
1. **Closing the intelligence gap between open and closed models** (09:43) — Fine-tuning near-frontier open models against specific business benchmarks enables enterprises to achieve performance that exceeds closed models.
1. **Balancing cloud infrastructure and internal artificial intelligence workstations** (12:51) — Organizations typically start by consuming external tokens for immediate results before adopting localized infrastructure and routers to reduce long-term costs.
1. **Deploying artificial intelligence in heavily regulated legacy enterprises** (15:45) — Utilizing confidential computing and open source models allows traditional companies to maintain strict data residency compliance while adopting new capabilities.
1. **Expanding open source resources and foundation models for developers** (18:14) — Providing open access to core libraries and diverse foundation models accelerates global market driven innovation across software and robotics.
1. **Promoting transparent infrastructure failure analysis through industry alliances** (20:35) — Cross-domain collaboration enables cybersecurity experts to trace agent actions and build more secure foundations after system escapes.
1. **Scaling consumer and enterprise communication through automated personal agents** (22:04) — Deploying millions of internal agents necessitates entirely new tools for large scale governance, observability, and automated inter-agent communication.
1. **Optimizing rack scale silicon infrastructure for specific model inference** (24:53) — Hardware providers must continually develop specialized kernels to maximize performance per watt for every newly released foundation model.
1. **Accelerating application development through open artificial intelligence harnesses** (26:44) — The rapid emergence of open source frameworks fundamentally shifts how developers build and manage transparent agentic systems.

## Related Moments

- [Autonomous agents and unusual startup pivots](https://www.wearedevelopers.com/videos/2144-developers-communities-in-san-francisco-wearedevelopers-live-from-corgi-cafe) (from "Developers Communities in San Francisco - WeAreDevelopers LIVE from Corgi Cafe")
- [Embedding generative AI in enterprise software platforms](https://www.wearedevelopers.com/videos/916-beyond-the-hype-real-world-ai-strategies-panel) (from "Beyond the Hype: Real-World AI Strategies Panel")
- [Designing agentic AI solutions for the enterprise](https://www.wearedevelopers.com/videos/1831-ai-for-enterprise-developers-dr-damir-dobric) (from "AI for Enterprise Developers - Dr. Damir Dobric")
- [Transitioning to AI-native engineering teams and workflows](https://www.wearedevelopers.com/videos/100508-legacy-as-a-launchpad-how-yahoo-mail-is-undergoing-a-product-and-engineering-transformation) (from "Legacy as a Launchpad: How Yahoo Mail is Undergoing a Product and Engineering Transformation")
- [Navigating competition and infrastructure in enterprise AI](https://www.wearedevelopers.com/videos/1098-decoding-trends-strategies-for-success-in-the-evolving-digital-domain) (from "Decoding Trends: Strategies for Success in the Evolving Digital Domain")
- [Security integration and AI skepticism in developer tooling](https://www.wearedevelopers.com/videos/1830-wearedevelopers-live-speculaitions) (from "WeAreDevelopers LIVE - SpeculAItions")

## Related Articles

- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)
- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)
- [WWC24 Talk - Scott Hanselman - AI: Superhero or Supervillain?](https://www.wearedevelopers.com/magazine/469-wwc24-talk-scott-hanselman-ai-superhero-or-supervillain)
- [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it)

## Related Jobs

- [Principal Software Engineer, AI Compute Infrastructure](https://www.wearedevelopers.com/jobs/ext/2847709-principal-software-engineer-ai-compute-infrastructure) at **ARM**
- [GPU Cluster Engineer, Systems & Platform](https://www.wearedevelopers.com/jobs/48411-gpu-cluster-engineer-systems-platform) at **Sciforium**
- [Staff Software Engineer, AI Compute Infrastructure](https://www.wearedevelopers.com/jobs/ext/3474377-staff-software-engineer-ai-compute-infrastructure) at **ARM**
- [Senior AI Developer](https://www.wearedevelopers.com/jobs/ext/2836034-senior-ai-developer) at **PwC**
- [Principal Software Engineer, AI Inference Cloud](https://www.wearedevelopers.com/jobs/ext/2854957-principal-software-engineer-ai-inference-cloud) at **ARM**
- [Partner Sales Director - AI Alliances - Model Providers](https://www.wearedevelopers.com/jobs/48429-partner-sales-director-ai-alliances-model-providers) at **Dynatrace**