TechPod Talks

TechPod Talks

Download on the App Store

TechPod Talks episodes

  • Platform Shifts, Skeuomorphic AI, and Change Management as the Real Constraint with Abe Ankumah

    Abe Ankumah has spent his career at the infrastructure layer, building through the internet, mobility, cloud, and now AI. He was part of the founding engineering team at Fulcrum Microsystems working on low latency switching silicon, co-founded and led Nyansa through its acquisition by VMware, then scaled the SD-WAN and SASE business past 18,000 customers as VP and GM before serving as Chief Product Officer at 1Password. He is now investing, advising, and, as Stephen gets him to admit on tape, building again. Stephen and Abe get into why adopting AI without rethinking your process just produces shiny things, what it means that every agent inside an enterprise is effectively a privileged user, and why he keeps going back to the entrepreneurial ecosystem on the African continent.

    Main Topics Covered:

    • Skeuomorphism in AI adoption: why filming a stage play instead of making a movie is exactly what most companies are doing with AI, and how speed without rewiring the process is lipstick on a pig
    • Agents as privileged users: the credential and identity problem Abe was building toward at 1Password, and what just in time access for a single task actually requires
    • What it takes to sell the company you built: distribution, strategic fit, and why the VMware acquisition was the start of something rather than the end of it
    • Brutal intellectual honesty and solving problems from the customer backwards, learned running product and business operations at Aruba Networks
    • Change management as the binding constraint on AI: new organizations can be AI native from day one, established ones have to bring people along deliberately
    • Talent is uniformly distributed, opportunity is not: education, entrepreneurship, and why Abe sees sub-Saharan Africa where Southeast Asia was 20 years ago

    Links & Resources

    • Connect with Abe Ankumah: LinkedIn
    • Connect with Stephen Koza: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    Companies & References:

    • Ashesi University Foundation, where Abe serves on the Board of Trustees
    • 1Password, where Abe was Chief Product Officer
    • Nyansa, the network analytics company Abe co-founded in 2013, acquired by VMware in 2020

    45 min
  • Product Market Fit, Secure by Design, and Betting on the Models with Daghan Altas

    Product Market Fit, Secure by Design, and Betting on the Models with Daghan Altas

    Daghan Altas is Head of Product at Semgrep, a code security company that helps engineering teams find and fix vulnerabilities in the code they ship. He's spent his career moving between product and sales, from creating the MX security appliance and SD-WAN product line at Cisco Meraki to building Semgrep's sales function as Chief Revenue Officer before returning to lead product. Stephen and Daghan, who worked together at Meraki, get into how to recognize real product market fit, why AI has turned code security into a race for defender velocity, and why he believes human code review is on its way out.

    Main Topics Covered:

    • Building the MX product line at Meraki: why most of product management comes down to calling customers, and his test for product market fit, a $100K deal closing without him in the room
    • Breaking the sales leader catch-22 at Semgrep: taking over sales before the company could attract a seasoned sales leader, and why the best product leaders have real empathy for sellers
    • Secure by design over vulnerability hunting: why X-raying a finished bridge is the wrong model for software security, and why the fix is more cultural than technical
    • Betting on the models: why a product built to patch a model's current weakness has a one-year lifespan, and why the real race in code security is closing the full loop from detection to remediation to proof of work
    • Where to start if you're behind on AI: clock eight hours a week, start with the omelet before the full Italian menu, and put your strongest AI people on one team instead of spreading them thin
    • Coding stopped being the bottleneck: why line-by-line human PR review is going away, and why humans should stay focused on architecture and threat modeling

    Links & Resources

    • Connect with Daghan Altas: LinkedIn
    • Connect with Stephen Koza: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    References:

    • OWASP Top 10 from the OWASP Foundation
    • Project Glasswing from Anthropic
    • Semgrep, where Daghan leads product

    44 min
  • Coding Stopped Being the Bottleneck. So What Replaced It?

    Back in April, Stephen Koza opened episode two with a claim: coding is not the bottleneck, it is everything else. Over the next ten episodes he put a version of that question to ten engineering leaders, and they did not agree with each other. This episode runs their answers side by side and lands on the one thing they had in common, which is that almost none of the answers were technical.

    Main Topics Covered:

    • Where the bottleneck actually went once code generation stopped being the constraint, from code review to deployment to go to market
    • Why story points are collapsing to a factor of one, and which agile roles can be handed to an agent
    • Request volume jumping 25x in a month, and why cost optimization is always the first work to get dropped
    • Token costs passing human costs, and the case for hiring people back
    • Why AI pilots stall on org structure rather than technology, and the argument for thinning vertically before flattening horizontally
    • The guests who reject the premise, from the limits of vibe coding to the junior engineer pipeline

    Guests featured in this episode:

    • Mark Sass, Episode 1: LinkedIn
    • Francisco Trindade, Episode 2: LinkedIn
    • Dr. Janet Sherlock, Episode 3: LinkedIn
    • Chris Robertson, Episode 4: LinkedIn
    • Meir Wasserman, Episode 5: LinkedIn
    • Robert Gonzalez, Episode 6: LinkedIn
    • Chris Harden, Episode 7: LinkedIn
    • Jose Gonzalez, Episode 8: LinkedIn
    • Tom Kershaw, Episode 9: LinkedIn
    • Patrick McKinney, Episode 10: LinkedIn

    Connect with Stephen Koza: LinkedIn

    Connect with EverOps: Website | LinkedIn

    References:

    • 2025 DORA State of AI-assisted Software Development
    • The full write up and checklist on the EverOps site

    36 min
  • Building to Understand, Storytelling, and AI Adoption Friction with Joel Cloralt

    Joel Cloralt is Director of Product at Smartcar, a connected vehicle API platform that helps developers build applications connecting to cars, from routing Uber EVs by battery level to smartly managing grid load when everyone plugs in at 5pm. He's spent 15 years building products across fintech, developer tooling, and connected vehicle tech, all while shipping his own apps on the side. Stephen and Joel get into what it actually takes to make a hard pivot work, why clarity beats charisma in leadership, and where AI is genuinely changing how engineers work and where it's just noise.

    Main Topics Covered:

    • Leading the B2B to B2C pivot at Vault and the decision-making process that made the Acorns acquisition possible, and how to move on 80% conviction when 100% isn't an option
    • Why Joel builds his own apps as a product leader: shipping a Japanese-learning app in Tokyo taught him the difference between what's impressive and what people actually use
    • Introvert leadership in practice: why the room doesn't need you to be an extrovert, it needs you to be clear, and how preparation becomes the whole advantage
    • Storytelling as an organizational tool: why ideas that travel through a company are wrapped in narrative, not bullet points, and how playing a customer interview video to an engineering org beats any slide deck
    • AI at Smartcar: building a developer-facing MCP, making documentation agent-friendly, and the real heavy lift of making sure what the agent responds with is actually accurate
    • The underrated skill of knowing when to stop: in product, the effort gets praised, but judgment about what not to build is rarer and worth more

    Links & Resources

    • Connect with Joel Cloralt: LinkedIn
    • Connect with Stephen Koza: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    Books & References:

    • Storyworthy by Matthew Dicks
    • The Pyramid Principle by Barbara Minto
    • The One Thing by Gary Keller

    47 min
  • Career Pivots, Authentic Leadership, and the Agentic Era with Stephen Koza

    TechPod Talks producer Lauren Dort turns the mic on host Stephen Koza, CEO of EverOps. Stephen traces his path from electrical engineering student at UT Austin to enterprise sales, through Silicon Valley startups and investing, to acquiring and running EverOps since February 2022. The conversation covers what authentic leadership actually looks like in practice, how to hire for ceiling rather than credentials, why most companies are failing to get ROI from AI, and what it will really take before production systems can run autonomously.

    Main Topics Covered

    • What the squiggly career path actually teaches you, and why putting too much pressure on early-career planning often works against you
    • How to maintain credibility with technical teams when you are not the most technical person in the room
    • Why caring is a leadership behavior, not a personality trait, and what it looks like when someone actually does it versus just says it
    • The difference between hard to work with and toxic, and why conflating the two is a costly leadership mistake
    • The rising star hiring profile: what to look for in candidates whose ceiling far exceeds their current credentials
    • Why most AI initiatives are not making it to production, and what companies need to get right about data and infrastructure before the agentic era delivers real returns

    Links & Resources

    • Connect with Stephen Koza: LinkedIn
    • Connect with Lauren Hochman Dort: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    References

    • Radical Candor by Kim Scott
    • What You Do Is Who You Are by Ben Horowitz

    45 min
  • Security, AI Adoption, and Human Judgment with Patrick McKinney

    Patrick McKinney, VP of Security at Invisible Technologies, has spent his career building security and compliance functions from the ground up, from Coinbase to Dropbox to Salesforce. He shares the playbook lessons that carried across each stop, why phase two never adds security, and how he's applying an AI native approach to security at Invisible while still keeping judgment and accountability firmly in human hands.

    Main Topics Covered:

    • Building compliance functions from scratch at Coinbase, Dropbox, and Salesforce, and the lesson that phase two never adds security
    • Standing up SOX and GDPR compliance ahead of Dropbox's 2018 IPO
    • Early FedRAMP work at Salesforce, including a dedicated authorized colo built entirely from scratch
    • Investing in an agentic SOC platform and running quality bake offs before adopting new AI security tools
    • Knowing when to shift security ownership from a CTO or head of IT to a dedicated security professional
    • How AI is reshaping entry level talent, and why judgment and accountability still belong to humans

    Links & Resources:

    • Connect with Patrick McKinney: LinkedIn
    • Connect with Stephen Koza: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    46 min
  • Prebid, Music Royalties, and Leading Through Transformation with Tom Kershaw

    Tom Kershaw, SVP and CTO at BMI, shares how a career spanning telecom, ad tech, travel, and now music rights taught him to learn new industries fast and lead through transformation. He breaks down the massive data challenge of tracking billions of music performances, the open source story behind Prebid that led to Google's antitrust trial, and why trying to stop AI in music is a losing battle.

    Main Topics Covered:

    • Crossing industries from telecom to ad tech to music rights, and why going deep on one thing (not everything) is the fastest way to learn a new business
    • The Prebid open source story: uniting competing ad tech companies against Google's walled gardens, and testifying in the DOJ antitrust trial
    • How BMI tracks billions of music performances across streaming, live venues, TikTok, and elevators to pay 1.5 million songwriters
    • Solving two data problems at once: high-complexity structured ownership data (SQL) and high-scale streaming counts (Kafka, Databricks, Spark)
    • Why AI in music can't be stopped, how songwriters already use it, and the ethical frameworks BMI is building around opt-in and attribution
    • Transformation leadership: why "focus" literally means "no," and the willingness to fail and course correct is the only way through

    Links:

    • Connect with Tom Kershaw: LinkedIn
    • Connect with Stephen Koza: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    47 min
  • From Six Sigma to AI: Turning Discovery Into Revenue with Jose Gonzalez

    Jose Gonzalez, VP of Product at Ticketmaster, shares how a career spanning the dot-com bubble, Johnson & Johnson, Yahoo, EA, Amazon, and a Techstars-backed AI startup shaped his approach to product leadership. He breaks down the power law behind media platforms, what it takes to scale teams under pressure, and why trying to put the AI genie back in the bottle is the wrong move.

    Main Topics Covered:

    • An unconventional path from web dev and Six Sigma process engineering at J&J to VP of Product at Ticketmaster
    • How power law dynamics drive media platforms, and why you can't build your business around hits alone
    • Co-founding an AI startup (Naria) through Techstars, and what the experience taught him about timing and product-market fit
    • Scaling teams fast: hiring for horsepower over perfect fit, and knowing when to lower the bar without dropping it
    • Why AI adoption mirrors the MP3/Kazaa/Spotify era, and the leaders who stop resisting will define what comes next
    • Using AI at Ticketmaster to connect fans with the right event at the right moment through smarter signals and discovery

    Links & Resources

    • Connect with Jose Gonzalez: LinkedIn
    • Connect with Stephen Koza: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    38 min
  • Systems Thinking, Shark Tank, and AI in Gaming with Chris Harden

    Chris Harden, Director of Production Verification and AI Liaison at Unity, shares how a career spanning theme parks, embedded software, EA Sports, and a startup that was featured on Shark Tank shaped his approach to building the systems around software. He breaks down the AI debate in gaming, how he used agents to write his latest book, and why communication is the most underrated leadership skill.

    Main Topics Covered:

    • A career built on systems thinking, from programming Epcot control systems to shipping NBA Live at EA Sports to co-founding a Shark Tank startup
    • Why the hardest part of shipping a triple-A game isn't the code, it's coordination and integration across hundreds of people
    • How Chris uses AI agents at Unity, including a jargon remover, a Scrum Master agent, and a multi-agent workflow for writing his latest book
    • The AI and gaming debate: where rapid prototyping wins, where human taste still matters, and why engineers are the orchestrators and verifiers
    • Communication as the most common leadership gap, from IC to executive, and why everyone should work in sales at least once

    Links & Resources

    • Connect with Chris Harden: LinkedIn | Website
    • Connect with Stephen Koza: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    Books Referenced:

    • Scale by Geoffrey Moore (systems thinking)
    • Crossing the Chasm by Geoffrey Moore
    • Little Robot, Big Dreams by Chris Harden (Shark Tank memoir)
    • Mastering Commitment by Chris Harden (building trust in teams)
    • The 4 Piece Operating Plan by Chris Harden (operating plans framework)

    43 min
  • AI-Native Engineering: From Coding to Orchestration with Robert Gonzalez

    Robert Gonzalez, VP of Engineering at SugarAI shares how he's transforming a 115-person engineering org into an AI-native operation, from building Scrum Master agents to rethinking how engineering value is measured. He also makes the case that in the age of AI, your product's durable moat isn't the UI, it's the context layer underneath it.

    Main Topics Covered:

    • A self-taught career path from webmaster to VP of Engineering, and the value of saying yes to risky projects
    • Leading a globally distributed engineering org across four continents using the "golden window" for collaboration
    • Why AI acceleration in engineering just moves the bottleneck, and why the whole business needs to keep pace
    • Building AI agents to automate Scrum Master duties, sprint metrics, and backward-looking estimation analysis
    • Why the context layer, not the UI, is the durable moat for SaaS companies in the AI era

    Links & Resources:

    • Connect with Robert Gonzalez: LinkedIn
    • Connect with Stephen Koza: LinkedIn
    • Connect with EverOps: Website | LinkedIn

    47 min

About TechPod Talks

From the publisher's feed

Are you a tech leader or executive dealing with the real pressures of running production at scale, building teams with limited bandwidth, and figuring out how to actually make AI work? These aren't hypothetical challenges, they're what's keeping engineering leaders up right now, and solid, practical advice is hard to find.