Utkarsh spent three years at Optum, the services arm of UnitedHealth Group, where he led the roadmap for a HIPAA-compliant machine learning platform that audited medical call reviews. The job was reading how claims conversations actually go, at scale: by the time he left, the platform had cut manual review work by 40 percent and raised audit accuracy by 15 percent. It is hard to spend three years inside healthcare operations and keep any illusions about what happens between a claim and a payment.
Then came three years at Amazon, first leading selection discovery across nine European marketplaces, then owning the selection and catalog quality roadmap for Gulf retail out of Dubai. That work runs on the same fuel as this site: customer complaints, defect data, and records that do not match, read until the pattern confesses. His systems automated the delisting of unreliable products and cut post-purchase defects by 10 percent.
He holds an MBA from the Indian Institute of Foreign Trade, Delhi, and an engineering degree from Pune. He now builds data warehouse and order-to-cash products at an e-commerce startup.
Two things he is not: a licensed insurance agent, and a seller of anything on this site. The Peril Desk never tells you what to buy. Every story is checked against the source ledger printed at the bottom of its page, every ranking copies named public data with the vintage stated, and when something here is wrong, the correction gets published and linked. That is the whole method: read the documents, show the receipts.