The code review problem nobody talks about
Pull requests pile up. Maintainers burn out. Subtle bugs slip through because the reviewer was on their fifth PR of the day and their eyes glazed over at line 247. Tests and linters catch the obvious stuff — but the 3-AM-wake-up-call bugs? Those need a second set of eyes that never gets tired.
David Parry thinks AI is that second set of eyes.
Who is David Parry
David is a Principal Architect at Qodo with decades of experience in enterprise architecture, Java, and building scalable systems. He is not a newcomer chasing the AI hype — he is a seasoned architect who has seen enough production incidents to know that code quality is not a nice-to-have, it is the difference between sleeping through the night and getting paged at 3 AM.
His focus: agent-driven developer productivity. Not AI that generates code for you to debug, but AI that reviews code alongside you — catching the patterns, edge cases, and architectural concerns that humans miss when they are context-switching between twelve repositories.
Free AI code reviews for open source
Here is the announcement that got my attention:
If you maintain an open-source project, Qodo will sponsor your code reviews — for free.
That is not a limited trial or a freemium bait. It is a commitment to making open-source software more secure and efficient by giving maintainers access to AI-powered code review agents that:
- Catch subtle bugs that linters miss
- Identify security vulnerabilities in pull requests
- Reduce review noise so maintainers can focus on what matters
- Increase confidence to merge and ship
For open-source maintainers who are already volunteering their time, this removes one of the biggest bottlenecks: the review queue. Head to qodo.ai to get started.
Agent MCPs: bridging AI and developer workflows
David is speaking at KubeAutoDay on March 23rd at the Amstel Boathouse in Amsterdam. His session, “Introduction to the Agent MCPs”, dives into how Model Context Protocols are becoming the bridge between AI agents, developer workflows, and system tools.
MCPs matter because they solve the integration problem. Having a smart AI agent is useless if it cannot access your codebase, understand your CI pipeline, read your Kubernetes manifests, or interact with your issue tracker. MCPs provide the standardized protocol layer that connects AI agents to the tools developers already use — without building custom integrations for every combination.
This is directly relevant to anyone building AI-powered developer platforms or thinking about how agentic AI fits into their engineering workflow.
Why David matters for the ecosystem
The intersection of AI and developer productivity is crowded with marketing. What makes David’s approach different:
- Decades of real architecture experience — he knows what production code actually needs, not what looks good in a demo
- Open-source first — offering free code reviews to OSS maintainers is a genuine investment in the ecosystem
- Enterprise Java roots — he understands the complexity of large-scale systems where AI code review has the highest impact
- Agent architecture focus — MCPs are the infrastructure layer that will determine whether AI agents become genuinely useful or remain glorified autocomplete
Catch David at KubeCon EU 2026
- 📅 KubeAutoDay — March 23rd, Amstel Boathouse, Amsterdam
- 🎤 Talk: “Introduction to the Agent MCPs”
- 🔗 Qodo: qodo.ai
- 🔗 LinkedIn: daviddryparry
This is part of my KubeCon Europe 2026 Leaders series. I will be at KubeCon EU 2026 speaking about Multi-tenant GPUs on Bare Metal OpenShift AI. Check out the full side events guide.
