The Forward Deployed Engineer

S.SHIMODA

AI works. The question is whether the team that bought it has ever sat at the workstation that's supposed to use it.
The Last Mile is a podcast about the hardest part of enterprise AI — the gap between a generalizable model and the messy reality of a specific customer's operations. Frameworks, war stories, and the economics of making deployments actually land.

Play

Open in

9 episodes
Avg 41 min
English

Episodes 9

Newest Oldest Longest

All9 Unplayed9 In progress0 Played0

July 23 · 40 min
The Outer Loop (Paving Gravel Roads Into Scalable Software)
<div> In this episode, we dive into Chapter 9 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda. We explore the "Outer Loop"—the strategic phase that begins the moment a deployment goes into production, ensuring your FDE function creates compounding platform leverage rather than slowly degrading into a high-end consulting business [1, 2]. Tune in as we break down the core analogy of the chapter: transforming messy, customer-specific code ("gravel roads") into generalized, reusable platform capabilities ("paved superhighways") [3, 4]. We explore the tactical frameworks and feedback mechanisms that make the outer loop actually work: <ul> <li>The Productization Decision: How the committee evaluates if a custom solution is worth paving based on reusability, differentiation, platform fit, and investment cost [5-7].</li> <li>Continuous Feedback Channels: Building vital links between FDE pods and the core platform team through engagement retrospectives, direct platform pull requests, sync meetings, and field rotations [8-10].</li> <li>The Platform Commit and Two-Team Handshake: Formalizing the agreement between field and platform engineers to safely migrate customers and retire old gravel roads [11-13].</li> <li>Outer Loop Failure Modes: How to avoid dangerous traps like "gravel-road accumulation" (where your team drowns in maintaining custom code) or "platform capture" (where the core product becomes bloated with one-off requests) [14, 15].</li> </ul> Finally, we reveal the four metrics every FDE leader must track to measure their platform leverage, ensuring that your next customer's deployment starts a step ahead, ultimately driving down costs and improving your gross margins [16, 17]. <hr> <h3>📚 Get the Book!</h3> If you are interested in these contents and would like to know more about scaling field intelligence and building a functional outer loop, please purchase Sho Shimoda's book on Amazon and tell others about it! Buy it here: <a href="https://www.amazon.com/dp/B0H2LJ1YM6" target="_blank">The Forward Deployed Engineer on Amazon</a> Thank you to our listeners for tuning in! Please follow, like, leave comments, and tell your friends to help spread the knowledge to the next era of software engineering.</div>
July 22 · 44 min
The Inner Loop (Prototype to Production)
In this episode, we dive into Chapter 8 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda. We explore the "Inner Loop"—the critical, iterative phase of an engagement that bridges the gap between the end of customer discovery and the first production milestone [1]. Tune in as we break down the tactical frameworks FDEs use to turn a generalizable platform into a customer-specific reality [2]: The DARE Framework: Mastering the tight cadence of Design, Activate, Realize, and Evolve to force continuous customer feedback on a one-to-two-week cycle [3-5]. Minimum Viable Architectures (MVAs): Why FDEs must fight the engineering instinct to over-build for the future, and instead ship deliberately under-built, easily replaceable architectures that solve the immediate workflow slice [6, 7]. Demo-Driven Development: Using rigidly scheduled demonstrations to actual users as a forcing function to discipline scope, ensure quality, and accelerate learning [8, 9]. The Hardening Phase: The crucial two-to-four weeks spent preparing the inner-loop prototype for actual production by addressing latency, reliability, observability, and strict enterprise security requirements [10-13]. We also discuss common inner-loop failure modes, such as scope drift and the "demo trap" [14, 15], and reveal exactly when you should—and shouldn't—invite the customer's own engineering team to help build the deployment [16, 17]. 📚 Get the Book! If you are interested in these contents and would like to know more about executing the perfect inner loop for enterprise AI, please purchase Sho Shimoda's book on Amazon and tell others about it! Buy it here: The Forward Deployed Engineer on Amazon Thank you to our listeners for tuning in! Please follow, like, leave comments, and tell your friends to help spread the knowledge of the next era of software engineering.
July 10 · 26 min
Customer Discovery and the Messy Reality (Solving the "Weird Tuesday" Problem)
In this episode, we dive into Chapter 7 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda [1, 2]. We explore the crucial first two weeks of an FDE engagement, known as the discovery phase, and why compressing this timeline to rush into building is a costly mistake that leads to failed deployments [2, 3]. Tune in as we break down the three concrete outputs every successful discovery phase must produce [3, 4]: The Stakeholder Map: Identifying sponsors, users, skeptics, and passengers, and using "snowball mapping" to find the people who can actually determine the engagement's outcome [4, 5]. The Workflow Inventory: Moving beyond static process diagrams to document real operational details like inputs, exceptions, handoff points, and metrics [6, 7]. The Eval-Customer Split: Making the critical business decision about which parts of a workflow the AI will handle automatically versus which parts will be routed to a human reviewer [8-10]. We also discuss the "Weird Tuesday" Problem—the atypical edge cases that happen on a random Tuesday, which standard AI demos miss, but which often contain the most operational value and risk [11, 12]. You will learn why FDEs use asynchronous interviews to uncover these hidden realities instead of relying solely on executive sponsors, who often don't know the actual floor operations [13-15]. 📚 Get the Book! If you are interested in these contents and would like to know more about mastering the customer discovery phase, please purchase Sho Shimoda's book on Amazon and tell others about it! Buy it here: The Forward Deployed Engineer on Amazon Thank you to our listeners for tuning in! Please follow, like, leave comments, and tell your friends to help spread the knowledge of the next era of software engineering.
July 9 · 41 min
The Soft Stack (Why Diplomacy is a Technical Skill)
In this episode, we unpack Chapter 6 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda [1, 2]. We explore why the FDE role is the single function in the AI industry where "soft skills" often outweigh technical skills as a predictor of long-term success [3]. We push back on the idea that soft skills are just vague, innate traits [4]. Instead, we reframe them as the "Soft Stack"—learnable technical disciplines that happen to be applied to people and conversations rather than to systems and code [4]. Tune in as we break down the critical frameworks every senior FDE must master to survive a deployment: Structured Problem-Solving (MECE): Using the "Mutually Exclusive, Collectively Exhaustive" framework to decompose ambiguous business problems into tractable engineering tasks without missing branches or duplicating work [5]. The Pyramid Principle: How to communicate with impatient executives by delivering your conclusion first, followed by grouped arguments, and saving the technical details for when they are requested [6, 7]. Translating Goals into PRDs: The under-taught skill of turning a vague executive aspiration (like "make us more efficient") into a precise Product Requirements Document [8, 9]. Managing the Skeptic: Why you should never route around the internal stakeholders who doubt your project, and how engaging them directly can turn an active opponent into your strongest sponsor [10, 11]. Stress Inoculation & The Diplomatic Register: How to maintain composure when deployments fail before board meetings, and how to communicate harsh technical realities without breaking the customer relationship [12-14]. 📚 Get the Book! If you are interested in these contents and would like to know more about mastering the strategic and diplomatic skills required for enterprise AI, please purchase Sho Shimoda's book on Amazon and tell others about it! Buy it here: The Forward Deployed Engineer on Amazon Thank you to our listeners for listening! Please follow, like, leave comments, and tell your friends to spread the knowledge to the next era of software engineering.
July 8 · 39 min
The AI and Agentic Frontier (Moving Beyond Chatbots)
In this episode, we dive into Chapter 5 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda [1]. We explore the rapidly shifting technological landscape of enterprise AI and why the skills that produced a competent chatbot in 2023 will produce a brittle, failing deployment in 2026 [2, 3]. We break down the critical shift from passive, conversational chatbots to active, autonomous "agentic" workflows [4]. Because enterprise executives don't want to chat with their data—they want systems that take action—Forward Deployed Engineers (FDEs) must now master a completely different set of architectural primitives [2, 4]. Tune in as we explore the essential capabilities every FDE must master on the modern AI frontier: Multi-Agent Orchestration: Moving away from single large language models to constellations of specialized agents (researchers, writers, reviewers, tool-users), and how to avoid the "cascading-confidence problem" where agents confidently pass bad data down the line [5, 6]. Evals as a First-Class Discipline: Why evaluations are the new unit tests. We discuss building golden datasets, capturing "weird Tuesday" edge cases, and using adversarial testing to ensure deployments don't go rogue [7-9]. Adversarial Testing & Red-Teaming: How to defend agentic systems against prompt injection, tool-call abuse, and data exfiltration before the customer's security team ever sees the deployment [10, 11]. Model-Agnostic Deployments: Why hard-coding your deployment to a single model is a massive mistake, and how to use routing abstractions to manage token costs and strict enterprise latency requirements [12-14]. Finally, we discuss what founders should actually look for when hiring for these skills, including why a candidate's "failure stories" and strong opinions on the AI frontier are their most valuable assets [15, 16]. 📚 Get the Book! If you are interested in these contents and would like to know more about mastering the AI and agentic layer for enterprise deployments, please purchase Sho Shimoda's book on Amazon and tell others about it! Buy it here: The Forward Deployed Engineer on Amazon Thank you to our listeners for tuning in! Please follow, like, leave comments, and tell your friends to help spread the knowledge of the next era of software engineering.
July 7 · 39 min
The Technical Bar (Why FDEs Are the Ultimate Escalation)
In this episode, we unpack Chapter 4 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda. We explore why an FDE must be, first and last, an engineer—and why their technical bar actually needs to be higher than a comparable platform engineer [1, 2]. Unlike Solutions Architects or Customer Success Managers, FDEs have no technical fallback. When an integration breaks at 3:00 AM before a demo, the FDE cannot escalate to the platform team because they are the escalation [3]. Tune in as we break down the four technical primitives every senior FDE must master to survive a deployment [4-6]: Production-Grade Programming: Writing reliable, maintainable code (typically in Python, TypeScript, or Go) within the messy, constrained realities and imperfect CI/CD tooling of a customer's environment [6-8]. Data Engineering: Managing the unglamorous data plumbing—SQL fluency, ETL design, and navigating inconsistent, undocumented real-world data—which actually makes up the bulk of enterprise AI work [9-11]. Cloud and Infrastructure: Operational fluency across AWS, GCP, and Azure, alongside containerization essentials like Docker and Kubernetes [12, 13]. Security and Compliance: Designing secure deployments from day one that meet strict enterprise requirements (like VPC Service Controls, encryption, and audit logs) rather than trying to retrofit them at the end [14, 15]. We also reveal the three "non-obvious" skills that separate the top tier of FDEs from the rest: debugging across opaque system boundaries, conducting code reviews under extreme uncertainty, and designing architectures based entirely on constraints the customer dictates [16-19]. Finally, we bust a major hiring myth by explaining why an FDE absolutely does not need to be an AI research scientist with a machine learning PhD [20]. 📚 Get the Book! If you are interested in these contents and would like to know more about mastering the technical skillset of an FDE, please purchase Sho Shimoda's book on Amazon and tell others about it! Buy it here: The Forward Deployed Engineer on Amazon Thank you to our listeners for tuning in! Please follow, like, leave comments, and tell your friends to help spread the knowledge of the next era of software engineering.
May 25 · 51 min
Where the FDE Sits in the Org (Avoiding the Classification Mistake)
In this episode, we unpack Chapter 3 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda. We explore why the first decision a company makes about its Forward Deployed Engineering (FDE) function is almost always wrong, a phenomenon Shimoda calls the "Classification Mistake" We break down the common traps of placing FDEs in the wrong reporting structure [3]. You will learn why putting FDEs under Sales turns them into pre-sale tech support, why putting them under Professional Services creates a toxic "consulting trap" driven by billable hours rather than platform leverage, and why placing them under the VP of Engineering starves them of the talent they need to thrive. Tune in as we discuss the three reasonable answers for where an FDE function actually belongs: Peer to Platform Engineering: Reporting directly to the CTO to balance generalized platform building with customer-specific deployments. Inside Product: Reporting to the Chief Product Officer to act as a direct research arm for the product roadmap. Standalone to the CEO: Treating FDEs as a primary go-to-market motion with executive-level authority. Finally, we explore how to actually build the team. We discuss the "Pod Structure" of cross-functional teams, why FDEs need to be staffed with a 50/50 mix of senior and junior talent, and why simply relabeling existing Sales Engineers or Customer Success Managers as FDEs is the most expensive mistake a founder can make. We also reveal why FDEs command a 25% to 40% compensation premium over traditional platform engineers. If you are interested in these contents and would like to know more about designing the perfect FDE organization, please purchase Sho Shimoda's book on Amazon and tell others about it. Buy it here: The Forward Deployed Engineer on Amazon Thank you listeners for tuning in! Please follow, like, leave comments, and tell your friends to help spread the knowledge of the next era of software engineering.
May 25 · 46 min
The Last-Mile Problem in Enterprise AI
In this episode, we dive into Chapter 2 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda. We explore the concept of the "last mile"—a term borrowed from telecommunications and logistics to describe the hardest, most expensive part of a deployment. You will discover why, contrary to popular belief, AI actually makes this last mile longer and explodes the hidden "integration tax" that traditional SaaS models left to the customer. We break down the critical shift from simple task-level AI to true AI-native operations. Because AI-native systems act autonomously rather than just giving recommendations, they require a complete workflow redesign and massively expand the political surface area of a deployment. Tune in as we explore the four frictions every AI deployment must overcome at the last mile: Data Friction: Navigating messy, inconsistent, and undocumented legacy data that polished vendor demos never show . Workflow Friction: Redesigning around the undocumented edge cases and tacit knowledge of the human operators . Political Friction: Managing the internal sponsors, skeptics, and saboteurs who can kill a project . Trust Friction: Earning the buy-in of the skeptical front-line users who have seen past tech projects fail. Finally, we discuss the core thesis of the chapter: in enterprise AI, the model itself is a commodity, and the redesigned workflow is the actual product. We reveal why the real last mile doesn't live in the API integration layer, but on the operating floor in the chair of the human agent. If you are interested in these contents and would like to know more about overcoming the hidden integration tax of enterprise AI, please purchase Sho Shimoda's book on Amazon and tell others about it. Buy it here: The Forward Deployed Engineer on Amazon Thank you to our listeners for tuning in! Please follow, like, leave comments, and tell your friends to help spread the knowledge of the next era of software engineering.
May 25 · 45 min
What is a Forward Deployed Engineer? (The Operator's Contradiction)
In this episode, we dive into Chapter 1 of The Forward Deployed Engineer: Architecting the Last Mile of Enterprise AI by Sho Shimoda. We explore the "Operator's Contradiction"—the paradox where operations-heavy businesses, like BPOs, are threatened by AI, yet possess the exact operational knowledge required to successfully deploy it. We trace the origins of the FDE role back to Palantir's "Delta" engineers, who originally pioneered the work of bridging the gap between a generalized platform and a customer's highly restricted, messy reality. You will discover why an FDE is not just another Software Engineer, Sales Engineer, or Solutions Architect. Instead, we break down the five "seats" every FDE must master to turn a demo into a deployment : The Technical Seat: Writing and shipping production-grade code against real customer data. The Operations Seat: Understanding the customer's workflow deeply enough to successfully redesign it. The Business Seat: Articulating the deployment's contribution to revenue, margin, or risk . The Diplomacy Seat: Navigating the customer's internal politics to build trust with executives and users alike. The Rendition Seat: Translating seamlessly between the operational floor, the executive suite, and the home-office platform engineers. Finally, we look at how the "new wave" of AI labs—including OpenAI, Anthropic, Runway, and Greptile—are reviving this critical role to act as the "missing face" of accountability when AI makes mistakes, conquering the last mile of enterprise AI integration . If you are interested in these concepts and want to master the definitive playbook for AI deployment, please purchase Sho Shimoda's book on Amazon Buy it here: The Forward Deployed Engineer on Amazon Thank you for listening, Please follow, like, leave comments, and share with your friends to spread the knowledge of the next era of software engineering.

Showing 1–9 of 9 episodes