SPSalman Parvezinmlsystems.hashnode.dev·1d ago · 5 min readClaimants Without Seats — the authority bug in our own ontologyOn ML Systems' Master Ledger a fact is not a value. It is a claim — a value, plus who made it, plus the grade of evidence behind it — and several claims sit on the same entry. The ledger decides which00
KPKPI Partnersinkpipartners.hashnode.dev·3d ago · 6 min readModeling Oracle EBS Manufacturing Data for Analytics: A Data Engineer's GuideOracle EBS manufacturing data is scattered across a normalized schema that was designed to process transactions correctly, not to answer analytical questions quickly. If you've been handed a request t00
YWYULUX Windows & Doorsinyuluxwindowdata.hashnode.dev·Sep 4 · 5 min readA practical data model for a window scheduleA window schedule often starts as a spreadsheet. It may later support design review, quotation, drawing coordination, and manufacturing checks. Problems appear when one cell tries to hold a dimension,00
KPKPI Partnersinkpipartners.hashnode.dev·Aug 25 · 7 min readHow to Model Service Analytics for Salesforce Across Cases and Service RequestsA scalable Service Analytics for Salesforce architecture extracts service data without disrupting operational workflows, retains important lifecycle events, separates Cases from Service Requests, and 00
NRNylah Reynardinhumanpen.hashnode.dev·Aug 19 · 6 min readAI Humanizer Limits Need a Schema, Not a RankingA comparison that puts 8,000 words, 10,000 characters, 125 words, and unlimited in one Limit column has a type problem. Those values do not describe the same boundary. One may belong to a paid plan, a10
SNSpecial Needs Care Networkinspcn.hashnode.dev·Aug 13 · 4 min readModelling a Provider Taxonomy When One Provider Belongs to Several CategoriesMost directory schemas begin the same way: a providers table with a provider_type column, an enum with two or three values, and a hub page per value. It is the right first move. It stops being right t00
SNSpecial Needs Care Networkinspcn.hashnode.dev·Aug 11 · 3 min readBuilding a City-First Event Directory: Location and Category as First-Class DataAn event directory looks simple from the outside: search for a venue, open a profile, and ask for a quote. The underlying product problem is more specific. An event search is usually constrained by se00
SNSpecial Needs Care Networkinspcn.hashnode.dev·Aug 11 · 3 min readA Query Contract for Local Provider SearchLocal provider search is often treated as a text-ranking problem. A user types a city and a service, the system returns a list of pages, and the product team tunes the copy. That model is incomplete. 00
ABAniruddha Banerjeeinaniruddhawrites.hashnode.dev·Aug 6 · 23 min readThe DISTKEY That Was Right, and Still WrongAuthor's Note This account is inspired by recurring patterns observed in enterprise Amazon Redshift environments. Table names, business context, and specific figures have been generalized to protect c00
AAihh浩inaicompanionnotes9899.hashnode.dev·Jul 30 · 7 min readPin, Correct, Expire, Forget: A User-Controlled Memory Model for AI CompanionsDisclosure: I work with the LumiChat team. This article describes a vendor-neutral memory architecture we use when thinking about long-running character conversations. It is not an independent product00
