A reliability layer that decomposes an LLM response into atomic claims, checks them against retrieved evidence, and returns a confidence score with the reasoning attached.
PRODUCT MANAGER. BUILDER. CURIOUS PERSON.Hi, I’m
Hi, I’m
Guhan.
I like building products at the intersection of AI, security, and real-world problems.
Welcome to my portfolio.
Take a look around ↓a few things I bring along
Curiosity
Clarity
Initiative
THE USUAL COMPANYHover over us. We’re friendly.
BASED IN PITTSBURGH · STUDYING AT CMUSome things I’m working on↓
SELECTED WORKA mix of building, learning, and figuring things out.
On my workbench.
A music discovery product that ranks on what a song sounds like rather than on how many people already found it.
A working harness for evaluating AI product changes — fixed task sets, explicit graders, and run-over-run comparison.
WHAT I CAN DOSo, what can I
So, what can I
actually do?
I have worked across AI, cybersecurity, enterprise products and healthcare. Different rooms, same job: find the real problem, build something you can argue with, then check whether it worked.
Five questions I usually get asked. The underlined bits have receipts.
Currently: M.S. Product Management
Carnegie Mellon University
Carnegie Mellon University
RECENTLY ATOrthoBerryDeloittePathsetter AIParadigmIT Cybersecurity
A FEW FAIR QUESTIONS
I go and look at the actual work before I have an opinion about it. At OrthoBerry that meant before I touched the redesign. At ParadigmIT it meant that paid for a new engineer.
Yes, and I would rather build than argue. At OrthoBerry I so the team had something real to react to, and outside work I .
This is the part most teams skip. Before Alfred went anywhere near a pilot I , and I designed the product to instead of guessing.
Once, properly. I with a 15+ person team, and .
Positioning is a product problem, so I treat it like one. At ParadigmIT I rebuilt IronKlad’s narrative, pricing and demo, and . I also .
Have something
in mind?
Write me a note ↗Always up for a good product conversation.