# Build versus buy A pack of 8 sessions from the MLOps Community YouTube channel, in the order to watch them. 7h 04m of video. Page: https://mlopstalks.com/packs/build-versus-buy The prototype works, but nobody budgeted for the years of upgrades behind it. A managed service solves most requirements, yet one critical integration would reshape the application. An open-source tool looks free until the team tries to operate it. Start by defining the capability you need, then count the work that remains after either a purchase or an internal build. The practical accounts that follow disagree for useful reasons: team size, existing architecture and the value of specialized engineering change the answer. Compare a deliberate internal build with buying into an established data environment. Finish by separating the application, model and hosting decisions for generative systems, and make future replacement part of the original choice. Historical comparisons supply questions to ask, not a current shopping list. ## This pack is for you if - You need to compare a platform proposal with the full cost of maintaining an internal alternative. - Your team can buy most of a capability but would still have to build its important integrations. - An existing tool no longer fits, and you want the next choice to be easier to replace. ## The talks, in order ### 1. Modern ML Stack is a Lie Mike Del Balso, Tecton & Joe Reis | 22:20 | MLOps Community Video: https://www.youtube.com/watch?v=jOI40sv6CsM Summary: https://mlopstalks.com/talks/modern-ml-stack-is-a-lie.md Why first: Del Balso and Reis separate the workflow a team needs from the infrastructure that runs it. Their 2021 discussion shows how the same requirement can acquire different product names. Write down the work and its constraints before comparing offerings, or the shortlist may contain tools solving different problems. ### 2. Build vs Buy an ML Platform Diego Oppenheimer, Algorithmia | 57:20 | MLOps Meetup Video: https://www.youtube.com/watch?v=1bHQE11Qq0k Summary: https://mlopstalks.com/talks/build-vs-buy-an-ml-platform.md Why second: Oppenheimer expands a working prediction API into the responsibilities behind a production platform. Supported environments multiply integration work, and upgrades and user support continue after launch. This provides the missing side of many price comparisons: the ongoing engineering commitment of the internal option. ### 3. Machine Learning at Reasonable Scale Jacopo Tagliabue, Coveo | 1:04:32 | MLOps Coffee Sessions Video: https://www.youtube.com/watch?v=qbAo8mo88Ic Summary: https://mlopstalks.com/talks/machine-learning-at-reasonable-scale.md Why here: Tagliabue argues for buying infrastructure when the distinctive work is understanding domain data. His smaller-scale recommender examples also challenge assumptions that distributed systems are necessary. Establish the workload you actually have before using an ambitious internal architecture as the baseline for a purchasing decision. ### 4. Building an ML Platform at SurveyMonkey Shubhi Jain, SurveyMonkey | 55:42 | MLOps Meetup Video: https://www.youtube.com/watch?v=oq1g4s2dUHE Summary: https://mlopstalks.com/talks/building-an-ml-platform-at-surveymonkey.md Why here: Jain supplies a concrete case for building: the team wrote down data, serving and management requirements, found mismatches with available products, and had the engineering capacity to proceed. His account balances the preceding advice by making architectural fit and staffing explicit. Treat the reported delivery gains as this team's experience, not a forecast for every build. ### 5. MLflow vs Kubeflow 2022 Byron Allen, Contino | 1:05:40 | MLOps Coffee Sessions Video: https://www.youtube.com/watch?v=9YcLBSqZNzE Summary: https://mlopstalks.com/talks/mlflow-vs-kubeflow-2022.md Why here: Allen and Pearse's 2022 comparison makes adoption cost visible through cluster access, setup work and dependence on another team. Their products have different scopes, so the useful lesson is how to evaluate a trial and avoid sunk-cost reasoning. Read the dated feature judgments as examples of questions to test again. ### 6. MLOps with Databricks Maria Vechtomova, Ahold Delhaize | Marvelous MLOps | 52:44 | MLOps Podcast Video: https://www.youtube.com/watch?v=Oa6qZPlOv3c Summary: https://mlopstalks.com/talks/mlops-with-databricks.md Why here: Vechtomova describes a 2025 choice shaped by an already approved vendor and an existing data environment. Her actual deployment still mixes services where serving requirements demand it. Buying a platform can reduce assembly work without requiring every workload to stay inside that platform. ### 7. Gen AI Buy vs Build, Commercial vs Open Source Ilona Logvinova, McKinsey & Mohamed Abusaid, QuantumBlack, AI by McKinsey & Nayur Khan, Goldman Sachs | 56:21 | AI in Production 2024 Video: https://www.youtube.com/watch?v=IpXZGXeuHt4 Summary: https://mlopstalks.com/talks/gen-ai-buy-vs-build-commercial-vs-open-source.md Why here: This panel separates buying a complete application, building around a model API and hosting an open model. Each choice leaves different integration and operating work with the team. Provider changes also require prompt versioning and testing, so an apparent model substitution can carry application migration costs. ### 8. MLOps Critiques Matthijs Brouns, Xccelerated.io | 49:44 | MLOps Coffee Sessions Video: https://www.youtube.com/watch?v=SS2_jQN3sG0 Summary: https://mlopstalks.com/talks/mlops-critiques.md Why last: Brouns evaluates tools by how easily they can be removed when requirements change. Clear interfaces, limited coupling and exportable monitoring data make that concern concrete. Finish the decision with a replacement plan and an account of what your team must keep operating, whichever option it chooses today.