Activity
3K followers
Varun Sureshkumar replied to a comment
3mo
David Weitzenfeld,JD, CPA That's a really elegant way to put it - guardrails with freedom inside them. That maps perfectly to how the best release governance works too.
You define the non-negotiables: compliance checkpoints, rollback criteria, go/no-go thresholds. But within those? You give the team room to adapt, problem-solve, and move.
The mistake I've seen most often is confusing the guardrails with the entire plan and end up over-constraining instead of protecting them.
Curious how you think about where to draw that line in AI systems - especially when the guardrails themselves need to evolve as the model learns?
Varun Sureshkumar replied to a comment
3mo
Jaspreet Sidhu That's a great point i.e., the human element is exactly what makes this analogy so interesting to me. Neural networks are literally designed around how humans learn, so it makes sense that the failure modes map across too.
What strikes me most is that in both cases. AI models and project plans - the overconfidence comes from training on 'clean' data. In real programs, the messy variables are the norm, not the exception.
Are you seeing AI/automation change how teams approach planning and delivery in your work?
Education
Licenses & Certifications
Honors & Awards
-
TIAA RST 2020 People Choice Awards
TIAA
Awarded in 3 Categories
* Unsung Hero
* Prized Partner
* Dream Team -
TIAA Financial Services (TFS) Technology 2019 Award
TIAA
Awarded for Outstanding performance in 2019 for delivering excellence and leading the RST Release Program Management Team at TIAA.
Languages
-
English
Full professional proficiency
-
Tamil
Full professional proficiency
-
Telegu
Professional working proficiency
-
Hindi
Professional working proficiency
-
French
Elementary proficiency
Recommendations received
11 people have recommended Varun
Join now to viewOther similar profiles
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content