NAN132: The AI-Augmented Engineer
Garrett Masters, a Senior Network Administrator with 11 years of experience, discusses his journey into AI-augmented network engineering through his YouTube channel GtalksTech. He explains how AI tools like Claude and Gemini enable network engineers to leverage automation without deep programming expertise, using practical examples like building MCP servers for network management.
Summary
Eric Cho interviews Garrett Masters about the intersection of network automation and AI. Garrett shares his unconventional path to content creation, explaining that he started making videos as a way to learn in public and teach others simultaneously—a practice he finds reinforces his own understanding. He initially started with 250 subscribers and has grown his channel through practical, lab-based content that viewers can follow along with using downloadable YAML files and GitHub repositories.
Garrett introduces his concept of "Legacy Larry"—the network engineer still manually logging into devices without automation infrastructure—as his primary audience. He emphasizes that many network professionals struggle to find time for traditional programming education, which is where AI becomes transformative. Rather than requiring deep Python knowledge, engineers can use AI tools to generate automation code while maintaining domain expertise to validate outputs.
The discussion covers Garrett's workflow using Gemini Deep Research for content ideation and Claude for code generation and lab building. He describes his approach to context engineering—providing AI with sufficient information to work independently rather than relying on elaborate prompt engineering. Garrett addresses security concerns by discussing how he built an MCP (Model Context Protocol) server with read-only access and automatic redaction of sensitive data, allowing network engineers to safely share configuration data with AI systems.
On prompt engineering, Garrett notes that it's becoming less critical as models improve, with context engineering and data access becoming more important. He advocates for giving AI tools the right information rather than over-specifying instructions. The conversation touches on choosing appropriate tools (Python with Netmiko versus Ansible), the importance of vendor documentation accessibility for AI training, and how public, open-source documentation enables better AI reasoning.
Garrett emphasizes learning by doing and using AI to understand AI itself—having it write prompts for itself, create learning plans, and generate handoff documentation. He recommends people start with major vendors like OpenAI or Claude, experiment with their specific tools, and just begin building rather than waiting for formal training. His core philosophy is providing tangible, high-quality content rather than volume, with every video accompanied by usable code and detailed documentation.
About this episode
Garrett Masters talks with Eric Chou about the concept of the AI-Augmented Engineer. Together they discuss how IT professionals can use Python and NetMiko to automate tasks and even connect AI directly to lab environments. Garrett also shares practical advice on prompt engineering, the importance of context management, and how engineers can safely adopt AI<a class="excerpt-read-more" href="https://packetpushers.net/podcasts/network-automation-nerds/nan132-the-ai-augmented-engineer/" title="ReadNAN132: The AI-Augmented Engineer">... Read more »</a>
Key Insights
- Garrett found that creating content for others reinforced his own learning better than studying alone, turning content creation into a learning mechanism rather than just broadcasting.
- Network engineers face constant expansion of responsibilities (security, SD-WAN, Zero Trust, etc.), making AI tools valuable for automating routine tasks and freeing time for strategic work.
- Building custom MCP servers allows network engineers to create guardrails and bounds for AI safety, such as implementing read-only access or automatic sensitive data redaction.
- Network engineers can effectively use AI to generate Python code for automation despite not having deep programming expertise, as long as they can validate outputs using their domain knowledge.
- Garrett argues that prompt engineering as a discipline is diminishing in importance as LLMs improve, with context engineering and data access becoming more critical for AI effectiveness.
- The choice of AI tools and programming languages should be dictated by what the engineer already knows (Python vs. Ansible, for example) rather than what the AI defaults to generating.
- Vendors that make documentation public and accessible enable AI to reason about their systems more effectively than those keeping documentation proprietary or behind paywalls.
- Garrett uses system prompts and project-level context in Cloud to avoid having to repeat information like 'this is about my home lab' with every query, demonstrating implicit vs. explicit context strategies.
- Testing code generated by AI for security vulnerabilities and running it through security tools is more practical than manually auditing code, especially for engineers unfamiliar with specific programming patterns.
- Garrett's workflow involves using Gemini for less critical work and Claude for more important work, demonstrating strategic tool selection based on use case importance rather than one-size-fits-all approaches.
- Network engineers can safely share sensitive configuration data with AI by using custom scripts that redact sensitive information locally before sending data to remote LLMs.
- Garrett intentionally avoids producing high-volume content daily or weekly, instead focusing on higher-quality, tangible content that viewers can immediately implement, viewing frequent posting as a missed opportunity for depth.
Topics
Transcript
We're sponsored today by Curvium, an industry-leading system integrator that offers strategic IT consulting, professional engagements, automation, and AI. Curvium takes the time to understand your infrastructure needs and how to best support your business objectives. Curvium skill professionals for your needs first, to understand your infrastructure needs and how to best support your business objectives. Curvium's skilled professionals put your needs first, from rapid designs to full-scale architectural planning, and from short-term project completion to multi-year support. Curvium provides the know-how and vendor connections you need to succeed. Find out more at curvium.com. While you're there, check out their secure campus network architecture blueprint. That's curvium.com. While you're there, check out their secure campus network architecture blueprint. That's curvium.com.…
Full transcript available for MurmurCast members
Sign Up to AccessMore from The Everything Feed - All Packet Pushers Pods
TCG085: Enterprise Salescraft with Drew Meli
Drew Meli, an enterprise sales executive with 16 years of experience, discusses best practices in B2B enterprise software sales, emphasizing transparency, stakeholder alignment, and collaborative deal progression rather than high-pressure tactics. The conversation covers the complex journey from initial customer interest through procurement, security reviews, and the emerging impact of AI on sales processes and organizational structures.
TNO071: The Network Team Is Drowning. Is AI the Life Raft? (Sponsored)
Rekha Shenoy and Irfan Kimji from Backbox discuss how the exponential growth of vulnerabilities (49,000 CVEs annually) has made manual network operations unsustainable, and how AI-powered automation can help network teams manage patches and security updates at scale while maintaining human control and oversight.
HN840: How to Make a Technology Buying Decision
Sean Morgan, a research director at Deloro Group, discusses how technology buying decisions should extend beyond engineering specifications to include business alignment, ROI calculations, and understanding total cost of ownership. Engineers must shift from viewing IT as a cost center to positioning it as a business enabler by connecting technical decisions to revenue impact and organizational objectives.
IPB207: Flying Blind: Monitoring Might Not See IPv6
The IPv6 Buzz hosts discuss critical gaps in IPv6 monitoring across enterprise networks, highlighting that many monitoring platforms lack IPv6 awareness, vendor parity, and advanced analytical capabilities. They emphasize that while basic IPv6 data ingestion has improved, sophisticated features like cross-protocol event correlation, extension header analysis, and device identity tracking remain significant industry challenges.
N4N063: Link Layer Discovery Protocol
Link Layer Discovery Protocol (LLDP) is a standardized Layer 2 protocol that enables network devices to announce information about themselves to directly connected neighbors, facilitating network topology discovery and device identification in multi-vendor environments. The protocol uses Ethernet frames with special multicast destination MAC addresses to ensure frames don't propagate beyond immediate neighbors, and includes mandatory TLVs (Type-Length-Values) like chassis ID, port ID, and TTL alongside optional ones for extended information.