Reader Reflections on Issue #4
I love hearing from readers, and Srikanth took the time to share his thoughts on Issue #4. His observations are thoughtful, nuanced, and I think capture something essential about how AI project management is evolving. With his permission, I’m sharing them here.
On Practicality and Relatability
Srikanth found Issue #4 to be “more relatable, pragmatic with specific pointers (e.g., vector enquires) that could be applied in concrete AI Dev contexts of different project types.”
This feedback means a lot to me because that’s exactly the goal of this newsletter to move beyond theory and give you actionable insights that translate directly to the work you’re doing. Whether you’re building RAG pipelines, fine-tuning models, or deploying agents, the principles should be practical enough to apply immediately.
On the Non-Linear Nature of AI Discovery
One observation that really resonated with me was Srikanth’s take on the visual lifecycle we included in that issue. He noted that it clearly illustrates how the discovery process is “NOT always being sequential but driven by the multi-dimensional vectors or lenses at given stage and time.”
He drew an interesting parallel to the spiral model in traditional software development—but with a critical difference:
“Here be it GO or NO GO… Its inherent nature is discovery cycles even after GO… hopefully spiralling upwards in terms of safe and sustainable value delivery.”
This is such a key point. Unlike traditional projects where discovery ends at deployment, AI projects demand continuous learning and adaptation. The “GO” decision isn’t an endpoint, it is a milestone in an ongoing journey. We deploy, we observe, we learn, we iterate. And hopefully, each cycle spirals upward, delivering more value and building more trust.
This aligns perfectly with what we’ve been discussing in this issue. Tools like Fiddler and Patronus aren’t just checkpoints they enable that continuous discovery cycle. Fiddler observes and intervenes in production. Patronus stress-tests and uncovers weaknesses pre-production. Together, they ensure each spiral upward is built on a solid foundation of observability and resilience.
On the Human Element
In a lighter vein, Srikanth added this observation:
“I think AI project management will demand more diligent human intelligence throughout the life cycle including deployment and maintenance from ALL stakeholders.”
I couldn’t agree more. For all the talk of automation and autonomous agents, human intelligence remains irreplaceable—perhaps even more critical as AI systems grow more complex. Deployment and maintenance aren’t “set it and forget it” phases. They demand ongoing vigilance, judgment, and collaboration across teams: data scientists, engineers, product managers, compliance officers, and business stakeholders.
Tools can monitor, govern, and stress-test. But humans provide the context, the judgment, and the ethical grounding that no algorithm can replicate.
A Big Thank You to Srikanth
Srikanth, thank you for taking the time to share your reflections. Your insights, particularly on the spiral nature of AI discovery and the enduring importance of human intelligence are exactly the kind of thoughtful dialogue I hope this newsletter fosters.