In this episode of The Technical Co-Founder Podcast, Lucas and Luna explore how a growing SaaS startup replaced brittle mock-based tests with test containers — spinning up real PostgreSQL, Redis, and S3-compatible storage in ephemeral Docker containers for integration tests. The result: test flakiness dropped by 80 percent, false positives nearly vanished, and the CI pipeline sped up by 40 percent despite running more real-world scenarios. We walk through the before-and-after architecture, the key decisions around container lifecycle management, and how the team convinced skeptical senior engineers to abandon mocks. Plus, a practical look at where test containers shine versus where classic unit tests still make sense.

#TestContainers #IntegrationTesting #SoftwareTesting #Docker #ContinuousIntegration #Mocking #DevOps #BackendEngineering #SoftwareArchitecture #CIOptimization #PostgreSQL #Redis #S3CompatibleStorage #TechPodcast #EngineeringCulture #FexingoBusiness #BusinessPodcast #TheTechnicalCoFounder

Keep every episode free: buymeacoffee.com/fexingo

Episode 64 of The Technical Co-Founder Podcast with Fexingo dives into continuous profiling — a technique that captures production CPU, memory, and I/O usage at the function level in real time. Lucas and Luna examine how a fintech startup called ClearPay used continuous profiling to identify a single hot loop in their transaction engine that was wasting millions of CPU cycles per second. By fixing that loop and deploying a culture of weekly profile reviews, ClearPay slashed their AWS bill by 40 percent — roughly $1.2 million per year — and improved p99 latency by 30 percent. The hosts break down the difference between profiling and observability, the tooling choices (Pyroscope and eBPF), and why this technique is still underused outside of infrastructure-heavy teams. They also discuss the tension between profiling overhead and production safety, and how ClearPay convinced skeptical engineers to add profiling to their CI/CD pipeline. A practical episode for anyone building or running backend systems at scale.

#ContinuousProfiling #CloudCostOptimization #FintechEngineering #ClearPay #Pyroscope #eBPF #ProductionEngineering #BackendPerformance #AWS #SoftwareEngineering #BusinessAndTechnology #TheTechnicalCoFounderPodcast #FexingoBusiness #BusinessPodcast #StartupEngineering #Observability #CPUProfiling #LatencyReduction

Keep every episode free: buymeacoffee.com/fexingo

On this episode of The Technical Co-Founder Podcast, Lucas and Luna explore how a fintech startup called LedgerSync uses blameless postmortems to turn outages into engineering culture wins. They walk through a specific incident where an API timeout cascade took down payment processing for 40 minutes—and how the postmortem process led to a 90 percent reduction in similar incidents over six months. The hosts discuss the tension between accountability and psychological safety, the exact script for running a blameless postmortem meeting, and why the 'five whys' technique works better when depersonalized. They also touch on how smaller startups can adopt the practice without a dedicated SRE team. A practical episode for engineering leaders and founders who want to build a culture of learning, not blame.

#BlamelessPostmortems #IncidentResponse #EngineeringCulture #PsychologicalSafety #FiveWhys #StartupEngineering #Fintech #LedgerSync #APIOutages #SiteReliabilityEngineering #RootCauseAnalysis #TechnicalDebt #TeamTrust #BusinessAndTechnology #FexingoBusiness #BusinessPodcast #EngineeringLeadership #PostmortemCulture

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Technical Co-Founder Podcast, Lucas and Luna dive into how a fast-growing fintech startup called FloatPay uses automated regression testing to ship code multiple times a day without breaking things. They walk through FloatPay's specific approach: a pipeline that runs a full regression suite in under three minutes using parallelized containers and a clever diff-based test selection algorithm. Lucas breaks down the numbers — how this cut their pre-deployment testing time from 45 minutes to under three — and explains why most teams over-invest in integration tests when unit tests would catch 80 percent of regressions faster. Luna challenges whether this approach works for early-stage startups with tiny engineering teams, and they explore how FloatPay built its system incrementally. By the end, you'll have a concrete framework for building a regression testing pipeline that lets you ship with confidence, whether you're a three-person team or a fifty-person engineering org.

#RegressionTesting #ContinuousDelivery #FloatPay #Fintech #SoftwareEngineering #TestingAutomation #DevOps #CICD #StartupEngineering #CodeQuality #TestSelection #ParallelTesting #UnitTesting #IntegrationTesting #ShippingFast #EngineeringCulture #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Technical Co-Founder Podcast, Lucas and Luna dive into a concrete case: how a mid-stage fintech startup uses distributed tracing to cut mean time to resolution from hours to minutes. They walk through the implementation of OpenTelemetry spans, how the team convinced skeptical engineers to instrument their code, and the surprising discovery that 40% of their production bugs were actually infrastructure issues misattributed to application code. Lucas explains why distributed tracing is different from logging and metrics, and why it's the missing piece for teams running microservices at scale. A must-listen for CTOs and engineering leaders wrestling with observability.

#DistributedTracing #OpenTelemetry #Microservices #Observability #Fintech #StartupEngineering #Debugging #CTO #EngineeringLeadership #ProductionBugs #Infrastructure #ApplicationPerformance #SoftwareArchitecture #DevOps #Business #Technology #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Episode 60 of The Technical Co-Founder Podcast digs into a single case: how a mid-stage SaaS startup called LogiFlow built an internal developer platform (IDP) to reduce cognitive load on its engineering team. Lucas and Luna walk through why IDPs are gaining traction as an alternative to complex microservice architectures, and how LogiFlow's platform cut onboarding time from two weeks to two days while reducing production incidents by 30 percent. They discuss the specific trade-offs — including the upfront investment in platform engineering versus the long-term productivity gains — and why a well-designed IDP can let engineers focus on product features instead of infrastructure plumbing. The conversation ties back to broader trends in developer experience and platform thinking, with concrete numbers and lessons for any startup wrestling with growing engineering complexity.

#InternalDeveloperPlatform #PlatformEngineering #DeveloperExperience #CognitiveLoad #LogiFlow #EngineeringProductivity #Onboarding #IncidentReduction #Backstage #SpotifyModel #TechDebt #DeveloperPortals #StartupEngineering #SaaS #Infrastructure #Microservices #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Technical Co-Founder Podcast, Lucas and Luna dive into how a B2B SaaS startup replaced three customer support tiers with a single semantic search layer. The case: a company called SupportAI (fictional) built a vector-search system that routes, resolves, and escalates queries automatically using embeddings. Lucas walks through the architecture: precomputed embeddings for FAQ and documentation, real-time query embedding, cosine similarity scoring, and a confidence threshold that escalates to human agents when similarity drops below 85 percent. Luna presses on the trade-offs: training data requirements, cold-start problems, and how the team handled edge cases like ambiguous queries. The conversation also touches on the operational impact — the startup cut average resolution time from 12 hours to 11 minutes and reduced support headcount by 60 percent. They close with a forward-looking question: if semantic search can handle tier-1 and tier-2 support, what happens to the role of the customer success manager? A focused, technical conversation for engineering founders and CTOs building AI-driven products.

#SemanticSearch #VectorSearch #Embeddings #CustomerSupport #AIEngineering #StartupArchitecture #CosineSimilarity #B2BSaas #SupportAI #QueryRouting #ResolutionTime #ColdStart #EdgeCases #BusinessAndTechnology #FexingoBusiness #BusinessPodcast #TechnicalCoFounder #EngineeringLeadership

Keep every episode free: buymeacoffee.com/fexingo

In episode 58 of The Technical Co-Founder Podcast, Lucas and Luna explore how a startup called Opsless uses platform engineering to let product teams deploy infrastructure without a dedicated ops team. They walk through Opsless's core approach: building an internal developer platform (IDP) that abstracts Kubernetes, CI/CD, and monitoring into self-service APIs. Lucas breaks down how Opsless reduced time-to-production from weeks to hours for new services, and how they measured success using developer satisfaction scores and deployment frequency rather than uptime alone. Luna questions the trade-offs, including the cost of building versus buying a platform and the risk of creating a new bottleneck. The episode also touches on how Opsless convinced skeptical engineers to adopt the platform by starting with a single high-friction workflow: database provisioning. Listeners will come away with a concrete understanding of platform engineering patterns, the metrics that matter, and the cultural shift required to make it stick.

#PlatformEngineering #InternalDeveloperPlatform #IDP #DevOps #Kubernetes #SelfService #DeveloperExperience #Productivity #Startup #TechnicalCoFounder #FexingoBusiness #BusinessPodcast #Opsless #DatabaseProvisioning #DeploymentFrequency #DeveloperSatisfaction #Infrastructure #SoftwareEngineering

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna explore how a fintech startup called PayShield uses chaos engineering to deliberately break its own infrastructure in production. They walk through a specific case: how PayShield simulated a regional AWS outage to test its failover logic, discovering a silent deadlock in its payment-processing queue. The episode covers the team's weekly 'break things' schedule, the tooling they built on top of Chaos Monkey, and how they quantify resilience with a metric called Mean Time to Recover (MTTR). Lucas explains why most companies avoid chaos engineering until it's too late, and Luna pushes back on the idea of breaking production systems without customer impact. The conversation ends with practical advice for startups that want to start small with fault injection in staging environments.

#ChaosEngineering #Resilience #SRE #PayShield #FaultInjection #MTTR #AWS #ProductionTesting #Fintech #Infrastructure #SiteReliability #EngineeringCulture #DevOps #Startup #TechnicalDebt #BusinessAndTechnology #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Episode 56 of The Technical Co-Founder Podcast explores how a fintech startup called LedgerLayer used observability—not just monitoring—to slash its incident response time from over an hour to under twelve minutes. Lucas and Luna walk through the specific tools and cultural shifts that made the difference: structured logging, distributed tracing with OpenTelemetry, and a practice called 'error budgets for on-call.' They discuss why most teams confuse observability with dashboards, how LedgerLayer's CTO convinced engineers to instrument their code proactively, and the surprising side effect: fewer production bugs overall. If you're building or operating software, this episode delivers a concrete playbook for making your systems actually observable—not just monitored.

#Observability #IncidentResponse #OpenTelemetry #StructuredLogging #DistributedTracing #LedgerLayer #Fintech #SRE #DevOps #ErrorBudgets #OnCall #ProductionBugs #EngineeringCulture #StartupEngineering #BusinessAndTechnology #FexingoBusiness #BusinessPodcast #TechnicalCoFounder

Keep every episode free: buymeacoffee.com/fexingo