About
I'm Shiven. I transferred from the University of Washington to Michigan to study Information Analysis — a degree that essentially doesn't have a clean description. It sits between computer science, data science, product design, and organizational behavior. That was fine with me. Most of the problems I find interesting don't have clean descriptions either.
My work sits in the gap between product and technical execution. Not quite pure PM, not pure engineering. I'm usually the person asking what's actually broken before anyone builds a solution — which sounds obvious but is rarer than it should be.
The problems I like best are the ones where the obvious answer doesn't work. Where someone already tried it and it didn't stick. Usually that means the real issue is a workflow nobody documented, data that doesn't mean what people think it does, incentives that push teams in the wrong direction, or two groups that need to coordinate and have no clean mechanism to do it.
How I think
I treat systems problems before I treat product problems. If the workflow is broken, adding a feature to the interface doesn't fix anything — it adds another layer on top of a broken foundation. I try to be specific about what's actually wrong: not “data quality issues” but “these three columns have schema drift and nobody owns the fix.” Not “unclear user needs” but “two different user types are running through the same flow and both are confused for different reasons.”
I've built enough things to have a calibrated sense of what's actually hard versus what just sounds hard in a planning meeting. That distinction changes how I scope and how I prioritize.
What I've built
Snowflake-integrated data profiling pipelines for enterprise AI readiness at Jade Global. Marketplace analytics dashboards tracking millions of daily ad auctions at Verve. ML classification models on 450+ clinical patient records at NIMHANS. Product prototypes through full sprints at MProduct. AI-powered tools for my own problems. The range isn't strategic — it's just what showed up when I was paying attention.
What I'm looking for
Environments where the problem is still being figured out. I do well in early-stage or ambiguous situations: give me a vague problem, access to users and data, and a team that needs someone to connect the pieces. The roles I'm targeting — Forward Deployed Engineer, AI Product, PM, Product Analyst — all have in common that the job involves navigating messy real-world systems, not just optimizing known processes. That's the part I find interesting.
Outside work
I work out consistently. I read a lot about markets, startups, and how companies actually make decisions versus how they say they do. I build small tools for problems I have — that's usually where the most useful ideas come from. I play chess poorly but consistently, mostly because I like thinking about long-term position and tradeoffs even when I execute them badly.