Hiring top-tier talent, leveraging open source models, and staying competitive in the age of AI w/ Benny Chen #267

Benny Chen, Co-Founder @ Fireworks AI, joins the show to discuss his founder journey and share valuable insights on navigating common founder / product dev challenges in today’s agent-first landscape. He and Jerry cover strategies for creating effective messaging, staying competitive in a crowded market space, hiring top-tier talent / what qualities to look for in high-performing engineers, navigating the cultural shift to managing agents, creating data flywheels & how this can help your customers, and more.   ABOUT BENNY CHEN As co-founder and early product architect, Benny Chen shaped Fireworks AI’s infrastructure strategy, spearheading the design of scalable systems to support high-throughput AI model serving. Benny’s contributions established the technical foundation for Fireworks AI’s robust and cloud-native architecture, which underpins its ability to meet enterprise demands. Formerly Meta’s Ads Infrastructure Lead, Benny optimized large-scale ad-serving pipelines and developed significant expertise in distributed systems and cloud infrastructure. He holds a B.S. in Computer Science from Stanford University, bringing both leadership and technical depth to the Fireworks AI management team.   Sinch is the communications infrastructure the AI era runs on. There's a layer of infrastructure behind every text, call, and login code your product sends, and it works exactly like plumbing: nobody thinks about it until it's the reason something broke. Most providers route through 4-6 intermediaries; Sinch connects in 1-2 hops, direct carrier relationships across 600+ connections, handling 900 billion interactions a year across 60+ countries. Routing, compliance, fraud prevention handled automatically rather than manually managed by the business sending the message! Sinch is the reliability layer underneath AI-driven customer communications; the infrastructure that determines whether an AI agent's output actually reaches a real person as a delivered text, connected call, or verified interaction. Check it out here!   SHOW NOTES: Moving from an early idea to a rocket ship (1:13) Insights on developing / communicating your core message as an early founder (2:40) Role of open source & inference @ Fireworks AI (4:24) How Firework AI’s company messaging evolved over time (6:07) Popular customization features today (7:25) Strategies for staying competitive in a crowded market (9:11) Defining the customer data flywheel & how it helps users (11:25) Common types of data that companies can collect to train their AI models (13:53) The customer’s next steps after creating a data flywheel (15:53) Benny’s perspective on acquiring engineering talent as a founder (16:43) Common traits shared by high-performing engineers @ Fireworks (19:05) How AI has altered which traits founders prioritize when hiring (20:45) Navigating the shift from engineering work to managing agents (22:40) Frameworks to ensure agents are doing the right thing (25:23) Aligning your metrics with the outcome you’re trying to follow (27:37) What a typical day looks like for Benny as a founder (28:31) Advice for founders / eng leaders looking to embrace AI adoption (29:19) Rapid fire questions (31:15)   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In this episode, Pooja Brown (Founder @ Inventry.ai) shares her insights on balancing being a founder & technologist, especially within the mid-market manufacturing industry. We cover why founders need to lead with curiosity as they seek out customer problems to solve, strategies for solving complex problems related to supply chain, and strategies for selling your products. Pooja also dissects important fundraising tactics, how to identify areas that AI tooling can enhance within your business, reading customer signals, and bolstering your engineering skills by leveling up business capabilities.   ABOUT POOJA BROWN Pooja Brown is a technology executive and founder focused on building AI-native platforms that power real-world operations across industries. She has led engineering at scale at companies like Stitch Fix and DocuSign, building systems that combine data, workflows, and machine learning to drive everything from personalization and supply chain to digital agreements used by hundreds of thousands of businesses.Her experience spans multiple verticals including retail, enterprise SaaS, education technology, and real estate, where she has consistently focused on embedding AI directly into core business systems rather than layering it on top. Pooja is currently the founder and CEO of Inventry.ai, where she is building autonomous AI agents that help mid-market manufacturers run procurement and supply chain operations more effectively. Across her career, she has focused on turning complex operational data into systems that don’t just generate insights, but actually drive decisions and execution.   Sinch is the communications infrastructure the AI era runs on. There's a layer of infrastructure behind every text, call, and login code your product sends, and it works exactly like plumbing: nobody thinks about it until it's the reason something broke. Most providers route through 4-6 intermediaries; Sinch connects in 1-2 hops, direct carrier relationships across 600+ connections, handling 900 billion interactions a year across 60+ countries. Routing, compliance, fraud prevention handled automatically rather than manually managed by the business sending the message! Sinch is the reliability layer underneath AI-driven customer communications; the infrastructure that determines whether an AI agent's output actually reaches a real person as a delivered text, connected call, or verified interaction. Check it out here!   SHOW NOTES: What shaped Pooja’s entrepreneurship journey & background (3:20) Looking for the right problem & leading with curiosity (6:47) Insights on solving problems related to supply chain (10:42) Building systems for chaos / complexity (12:46) Adopting a beginner’s mindset when solving complex problems (15:40) Dissecting fundraising strategies & decision making (19:28) Emerging business patterns that eng leaders need to capitalize on (26:32) Understanding how customers make decisions on what products to adopt (29:09) Selling strategies for the mid-market manufacturing industry (37:08) Integrating AI tools to augment current business capabilities (41:03) Engineering skills that enhance sales processes (44:43) Communication frameworks when working with customers (47:21) Rapid fire questions (50:06)   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In this episode, Sergiy Nesterenko (CEO @ Quilter) joins the pod to share how he turned a personal technical challenge into his own company, even before the market caught up. He shares his unique journey from SpaceX to Quilter and addresses common eng leader / founder challenges, including taking critical feedback to solve tough technical problems, knowing when your most important role is educating the market / customers, and choosing the right customers. Sergiy and Jerry also dissect insights into principles around building circuit boards, and common misconceptions around hiring top-tier talent, finding investors, and procuring customers as an early-stage company.   ABOUTSERGIY NESTERENKO Sergiy Nesterenko is the Founder and CEO of Quilter, the first physics-driven AI platform that fully automates PCB layout design. Under his leadership, Quilter has pioneered autonomous board design – creating the missing automation layer for hardware and compressing weeks-long layout cycles into hours. Backed by Index, Benchmark, Coatue, and Tony Fadell, Quilter is redefining what's possible in hardware development. Previously, Sergiy spent five years at SpaceX as a Senior Radiation Effects Engineer, building electronics for the Falcon 9 and Falcon Heavy.   SHOW NOTES: Introducing Sergiy Nesterenko & the inspiration behind Quilter (1:43) Why now is the right time to tackle this technical challenge (3:37) The technology landscape during Quilter’s early days (6:05) Sergiy’s fundraising journey & key takeaways (8:26) The importance of adopting an evolutionary mindset (10:52) Early problems that Quilter aimed to solve (13:08) Navigating the balance between technical problems & consumer readiness (15:59) How negative feedback can lead to solving a key problem (18:34) Knowing which feedback to incorporate into product iterations (20:08) Lessons learned from choosing the wrong customers (27:33) First principles approach to circuit board building (29:31) Signals that it’s the right time to tackle a hard technical challenge (34:04) Advantages to solving really hard problems & tackling entrepreneurship (36:32) Common misconceptions surrounding early phase companies (38:17) Sergiy’s experience working with Eric @ Benchmark (39:46) Why investors like founders who spend time in the “idea maze” (42:19) Strategies for convincing high-profile talent to join your org early on (43:22) Rapid fire questions (45:10)   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Tomas Reimers, Co-Founder & CPO @ Graphite, reveals his journey to becoming a founder and some of the major moments from Graphite’s product development story. He dissects frameworks for determining when it’s time to quit your full-time job to pursue becoming a founder and analyzing the costs vs. benefits of that decision. We cover how to decide what customer problems to solve, understanding how software development trends impact what you’re building, and product market fit and selling considerations. Lastly, Tomas shares valuable strategies when it comes to selling to developers.   ABOUT TOMAS REIMERS Tomas Reimers is the CPO and co-founder of Graphite, the a16z and Anthropic-backed AI code review platform that Cursor acquired in December 2025. Previously, he was an engineer at Facebook. Passionate about advancing developer velocity, Tomas holds a BS in Computer Science from Harvard University.   SHOW NOTES: The origin story behind Graphite (1:05) Signs that it’s time to quit your job / found a company (4:03) How Tomas & his co-founders knew it was time to break up with their jobs (5:40) Understand what leaving your job will mean career-growth wise (7:33) Embracing regret minimization (9:28) Pros & cons of developing in blue ocean vs. red ocean areas as a founder (11:49) Strategies for determining what software to build / how is software dev evolving (14:00) Tomas’s framework for determining bets (15:50) Recommendations for selling as a founder (17:52) Identifying if sales isn’t working or if product market fit is off (20:35) Using customer insights to experiment & make a PMF pivot (23:05) Why developers don’t want to be sold to, they want to be taught (24:42) How to apply teaching vs. selling into early GTM strategy (27:20) Utilizing demonstrations to gain developer buy in & trust (29:24) What Tomas is looking forward to in Graphite’s future / software dev in general (32:04) Rapid fire questions (33:34)   LINKS AND RESOURCES The Mom Test - Rob Fitzpatrick’s quick, practical guide that will save you time, money, and heartbreak. Founding Sales - The Startup Sales Handbook - rhe distillation of Pete Kazanjy’s experience at his first software startup, TalentBin going from a founder with a product marketing and product management background to early sales guy, early sales manager, and eventual post-acquisition sales leader at Monster Worldwide in addition to his later experience founding Modern Sales, the nation’s largest sales operations, enablement, and leadership community. Thinking in Bets: Making Smarter Decisions When You Don't Have All the Facts - Annie Duke, a former World Series of Poker champion turned business consultant, draws on examples from business, sports, politics, and (of course) poker to share tools anyone can use to embrace uncertainty and make better decisions. This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Vivek Raghunathan, SVP of Engineering @ Snowflake, joins the Engineering Leadership Community Podcast to discuss all things AI – most importantly, how Snowflake is utilizing AI internally to iterate faster with smaller, focused teams. Vivek shares how AI and lower coding costs ultimately help Snowflake implement tighter feedback loops between customer & eng teams to speed up product development rollout and how Snowflake empowers their orgs to upskill when it comes to AI. Jerry & Vivek also break down what quality leadership looks like across the board and the role of AI in shaping today’s eng leaders.   ABOUT VIVEK RAGHUNATHAN Vivek Raghunathan is Snowflake's SVP of Engineering. Before joining Snowflake, Vivek co-founded Neeva, one of the first AI-native search engines for consumers, which was acquired by Snowflake in 2023. Prior to that, he spent more than a decade at Google, where he led YouTube monetization and helped launch Google Now.   ABOUT SNOWFLAKE Snowflake is the data and AI platform that many of the world's largest organizations rely on to power their most critical operations. Whether a bank is approving a loan, a retailer is optimizing its supply chain, or a hospital is bringing together patient data, Snowflake helps organizations unify data, apply AI, and make faster, more informed decisions. Today, over 13,900 companies rely on Snowflake to turn their data into business value, with enterprise-grade security, governance, and scale built-in.   SHOW NOTES: Where Snowflake is on their AI adoption journey (0:57) Using AI to help eng orgs meet maximum velocity (3:04) Defining traits of engineers who successfully leverage agentic tools (5:55) The benefit of cross collaborative skillsets for eng leaders (8:53) How AI is helping orgs test infrastructure rewrites effectively & less expensively (10:09) Discovering AI patterns @ Snowflake (11:44) Why Snowflake gives teams a week off to learn the necessary AI tools (16:37) Compressing the product dev cycle with the help of AI (18:28) Simplify eng teams / encourage fluidity of roles (19:49) Vivek’s perspective on leveraging AI to best support customer needs (22:45) Building control planes to support enterprises (24:30) Methods for consolidating data / making it accessible to the entire org (27:18) How AI shapes what a quality leader looks like (30:39) The role of empathy in leadership (33:49) Rapid fire questions (35:08)   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Russ d’Sa (CEO & Co-founder @ LiveKit) joins the show to deconstruct the "Product Paradigm Shift" toward voice-driven interfaces and agent-centric UX . We dive into LiveKit’s high-stakes scaling lessons: from powering OpenAI and Character AI’s voice mode, how they navigated real time bottlenecks to hit the next level of scale, the architectural necessity of a multi-cloud strategy, and the foundations of a co-founder relationships that can effectively blend engineering & business strategy.   ABOUT RUSS D’SA Russ is a startup vet who founded his first company in the 2007 YC batch and was the 2nd frontend engineer hired at Twitter, Russ d'Sa now leads voice AI unicorn LiveKit. They're the backbone of ChatGPT Voice Mode, Salesforce Agentforce, Grok, and roughly 30% of US 911 calls.   ABOUT LIVEKIT LiveKit is an open source framework and cloud platform for building voice, video, and physical AI agents. It provides the tools you need to build agents that interact with users in realtime over audio, video, and data streams. Agents run on the LiveKit server, which supplies the low-latency infrastructure (including transport, routing, synchronization, and session management) built on a production-grade WebRTC stack. This architecture enables reliable and performant agent workloads.   SHOW NOTES: The product paradigm shift toward voice-driven apps and natural human-computer interfaces (2:44) Voice-apps in practice: How these trends impact the strategy of product building today (5:32) Early adopters: Why legacy industries like healthcare use voice AI (7:55) Reevaluating and building product experiences optimized for AI agents (12:52) How AI trends will impact roadmaps and Go To Market (18:16) The origin of LiveKit: Building real-time infra for the pandemic (21:07) The OpenAI moment: Powering the fastest-growing consumer app (23:48) Scaling with OpenAI: Navigating the challenges of balancing time-to-market with system design (25:39) The Character AI outage: Solving cross-continental state sync and hitting the next level of scale (29:00) The problem: When telemetry breaks first: Managing analytics and logging for millions of concurrent AI sessions (32:04) Architecting for resilience: Multi-cloud from day one and why treating infra as a utility matters (33:22) Co-founder dynamics: Blending engineering strategy with business outcomes (37:15) Rapid Fire Questions (40:51)   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In this episode, Patrick discusses what it means to build an empowered career & explore creative career portfolios with Jean Hsu (VP of Engineering @ Range) and Cate Huston (author of The Engineering Leader and fractional CTO @ Twill). Both share their unique engineering leadership journeys & how they built creative career paths through exploration & finding room for optionality. We dissect the identity crisis that eng leaders face – whether they are ICs or managers – and how to navigate the tension between individual & team productivity, especially taking into consideration AI. Lastly, Jean and Cate share insights on letting go of societal norms, unique ways to expand your work, taking on bets, and incorporating your values into your career.   ABOUT JEAN HSU Jean is a builder, writer, coach, and fractional VPE at Circuit & Chisel. She was previously in leadership roles at Pulse, Medium, and Range, and also built out a leadership development company focused on engineers. She lives in Berkeley with her partner and three kids.   ABOUT CATE HUSTON Cate is the author of The Engineering Leader, fractional CTO at Twill, and engineering leadership coach. She was previously in leadership roles at DuckDuckGo and Automattic, and an advisor at Glowforge. She has been all over the world, but now lives in Ireland.   Check out DRI Your Career today and use promo code “ELCPODCAST” for 15% off any of Cate and Jean’s three courses!   SHOW NOTES: What it means for creative career paths to become the norm (1:42) Navigating the tension between individual vs. team productivity (3:34) What an empowered career looked like in Jean’s leadership journey (5:00) Cate’s decision to craft her own career narrative (10:46) Redefining work-life balance (12:54) How to cultivate time to explore future projects & create room for optionality (15:59) Why it can be challenging to find the space / time to experiment (19:17) Let go of “societal shoulds” (23:37) Frameworks for building out your career portfolio (28:43) Unique ways to expand the type of work you can perform (30:23) Using AI tools to help orient your career & incorporate your values (34:35) Thinking about your career portfolio as bets (40:14) Final thoughts on what it means to build an empowered career (44:47) Rapid fire questions (46:13)   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In this episode, we’re joined by Eric Ries, creator of The Lean Startup, to discuss insights from his latest book, Incorruptible: Why Good Companies Go Bad… and How Great Companies Stay Great. Eric shares what inspired him to write the book and why we need to move beyond and redefine what true profit looks like. He shares the history behind businesses transitioning from serving public interests to shareholder primacy and why leaving behind a people-first business approach can actually reduce profitability. Additionally, Eric discusses financial gravity, the “harder is easier” principle, and how these practices connect to AI & current engineering leadership challenges.   ABOUT ERIC RIES Over the last two decades, Eric Ries’s ideas about continuous innovation, long-term thinking, governance, and market reform have reshaped company building and management practices. He is the creator of the Lean Startup method, and the author of the New York Times bestseller The Lean Startup; The Leader’s Guide; and The Startup Way. As a founder, he has put his own ideas into practice with The Long-Term Stock Exchange (LTSE); Answer.AI, an AI R&D lab; Virgil, a legal services startup; and IMVU. On The Eric Ries Show, he talks with world-class technologists, thought leaders, and executives building for the long-term. He lives in the San Francisco Bay Area with his wife and three children.   Unblocked: The context engine your coding agents are missing. Give your coding agents the context your best engineers have. Your agents can read code, but they don’t know how your team works. Rules and MCPs give access to information but not understanding. That’s why you still have to tell them where to look and what to look for. Unblocked gives your agents the history, conventions, and decisions behind your code so they generate mergeable output without the back and forth. It automatically surfaces the right context for every task, so agents stay on track without the set up tax or the correction loops. getunblocked.com/elc   SHOW NOTES: The inspiration behind Eric’s new book Incorruptible (5:22) What it means to redefine profit (8:03) Understanding profit considerations like externality, ethics, and inputs (10:44) Why human life / value can never be an input factor of production (12:31) The history behind business practices benefitting the public (15:00) When businesses transitioned to shareholder primacy over public interest (17:16) Navigating the tension between mission vs. fiduciary responsibility (21:01) The role of financial gravity & shareholder primacy in the Silicon Valley bank story (25:04) Using Eric’s book to build a mission-driven roadmap (29:12) How committing to a principled way of business can drive profitability (31:15) An example of the principle “harder is easier” (33:40) How this connects to AI & emerging eng leadership challenges (36:53)   LINKS AND RESOURCES Incorruptible: Why Good Companies Go Bad and How Great Companies Stay Great - Drawing on two decades of work with founders, CEOs, and investors, best-selling author Eric Ries reveals the forces that make companies vulnerable to destruction from within and without. Then he offers solutions that safeguard against them for the long-term. Incorruptible is the blueprint for companies that will prosper and endure without losing their soul. Its lessons and tools are designed to help founders, executives, investors, and citizens of all kinds build organizations – and a society – truly aligned with human flourishing. https://news.theleanstartup.com/ - Eric’s newsletter with ideas about how and why to build companies focused on human flourishing — and stories of the people who are doing it. The Eric Ries Show - Founder, entrepreneur, and best-selling author of The Lean Startup Eric Ries discusses how to build profitable companies for the long-term benefit of society. Ries talks with world-class technologists, thought leaders, executives, and others working to create a new ecosystem of trustworthy organizations with limitless potential for growth and a deep commitment to purpose. Together, they uncover the tools and methods to ensure the next generation of companies are designed to maximize human flourishing for generations.   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In this episode, Geddes Munson (SVP of Engineering @ Affirm) joins us to discuss operational / engineering excellence, scaling, and AI-native transformation! We explore Affirm’s approach to operational and engineering excellence and how a 2024 outage became a turning point in refining that focus. We deconstruct “AI retooling week”, the internal tools it inspired (including an incident tracing system), how the AI-native transition is impacting operational / engineering excellence, and how to connect these projects to business goals. Plus, we take a look at their early work building in agentic commerce, infrastructure decisions they made years ago setting them up for success now, how they’re thinking about designing for agent-first experiences.   ABOUT GEDDES MUNSON Geddes Munson serves as Affirm’s SVP, Engineering. Previously, Geddes held several engineering leadership roles at Affirm, including oversight of the merchant engineering group, where he was responsible for the development of Affirm’s solutions for key partners including Amazon, Shopify and Walmart. Prior to Affirm, Geddes held various technical leadership roles at rapidly growing startups including Mixpanel, SingleStore and EasyPost. He received his B.A. from Haverford College, where he started the Linux club on campus. Geddes lives in New Jersey with his wife and three children.   Unblocked: The context engine your coding agents are missing. Give your coding agents the context your best engineers have. Your agents can read code, but they don’t know how your team works. Rules and MCPs give access to information but not understanding. That’s why you still have to tell them where to look and what to look for. Unblocked gives your agents the history, conventions, and decisions behind your code so they generate mergeable output without the back and forth. It automatically surfaces the right context for every task, so agents stay on track without the set up tax or the correction loops. getunblocked.com/elc   SHOW NOTES: Defining operational excellence & what it looks like @ Affirm (4:36) Understand why your company / product matters to your customers (8:11) Key pivot points around engineering excellence @ Affirm (11:10) Creating a genuine culture change of operational / engineering excellence (14:27) Adopting agentic models @ Affirm (16:30) Navigating the balance between transformation, safety & reliability (18:30) Affirm’s AI retooling week & hackathon setup (20:57) How the hackathon helped quickly change the company culture (23:15) Ensuring your practices serve your overall organizational vision & goals (26:11) Insights on scaling & increasing CICD investment @ Affirm (28:28) Approaches to building agentic commerce products (30:11) Strategies for building an agent-first experience (33:33) Bridging the gap between engineering & business goals / outcomes (35:44) Rapid fire questions (38:46)   LINKS AND RESOURCES 1929: Inside the Greatest Crash in History – and How It Shattered a Nation - New York Times bestselling author Andrew Ross Sorkin takes readers inside the chaos of the crash, behind the scenes of a raging battle between Wall Street and Washington and the larger-than-life characters whose ambition and naivete in an endless boom led to disaster. The dizzying highs and brutal lows of this era eerily mirror today’s world—where markets soar, political tensions mount, and the fight over financial influence plays out once again. Delivering Happiness: A Path to Profits, Passion, and Purpose - a best-selling 2010 memoir by former Zappos CEO Tony Hsieh detailing his entrepreneurial journey and outlines his core philosophy: building a phenomenal corporate culture and focusing on the happiness of employees and customers ultimately drives long-term profits and business success.   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Andrew McNamara, Director of Applied Machine Learning @ Shopify, joins the ELC podcast to share insights on building agentic platforms at scale, like Sidekick, that must keep reliability for its users at the forefront. Andrew describes the building philosophy behind Shopify and what it means to cultivate a culture of prototype-first while prioritizing hiring early-stage talent. We cover Sidekick’s development journey and how user feedback impacted its product vision, why evaluation is so important for determining ground truth sets, and the benefit of user-driven use cases. Andrew also dissects how they went about making product design decisions, such as building proactive agents and identifying subagent specializations.   ABOUT ANDREW MCNAMARA Andrew McNamara is Director of Applied Machine Learning at Shopify, where he leads the team behind Shopify Sidekick, an AI co-founder that gives merchants access to the e-commerce expertise they need to run and grow their business. With 16 years of experience building AI assistants, he brings a rare combination of applied research depth and production-scale thinking to some of the hardest problems in AI: getting systems to work reliably for people who depend on them. Andrew's work pushes Shopify to measure AI quality by whether it achieves what the user set out to do, a core standard in building AI that merchants trust. Outside Shopify, he runs Setting North, a small Canadian maple syrup brand built on the same platform he helps make for everyone else.   Unblocked: The context engine your coding agents are missing. Give your coding agents the context your best engineers have. Your agents can read code, but they don’t know how your team works. Rules and MCPs give access to information but not understanding. That’s why you still have to tell them where to look and what to look for. Unblocked gives your agents the history, conventions, and decisions behind your code so they generate mergeable output without the back and forth. It automatically surfaces the right context for every task, so agents stay on track without the set up tax or the correction loops. getunblocked.com/elc   SHOW NOTES: How Shopify utilizes reflexive AI & Andrew’s building philosophy (2:38) Developing a prototype-first company culture (5:07) Andrew’s reflections on building AI-enabled projects like Sidekick at scale (7:25) Translating customer surveys into Sidekick’s product vision (9:34) Key inflection points while scaling out Sidekick (11:23) Strategies for evaluation / building a ground truth set (13:26) Analyzing the good & bad within ground truth sets (15:27) Shopify’s system openness model to drive user-discovered use cases (17:47) How subagents fit into the Sidekick’s model (19:55) Prioritization conversations around subagent specializations (23:06) Designing an agent with high-impact prompt optimization (27:22) Considerations for building highly reliable systems (29:40) Andrew’s perspective on latency (31:24) Rapid fire questions (33:49)   LINKS AND RESOURCES Cradle - a New York Times best-selling series from Will Wight following a character's growth as he goes from one of the weakest users of his world's magic to among the strongest. The series features an original magic system inspired by Chinese cultivation and martial arts novels, with a heavy emphasis on anime-style super-powered battles.   This episode wouldn’t have been possible without the help of our incredible production team: Patrick Gallagher - Producer & Co-Host Jerry Li - Co-Host Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/ Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/ Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/ Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.