Data Infrastructure
8 technical briefings from 8 companies building data infrastructure infrastructure.
Amelia Niemi, co-founder and CTO of Ntropy, walks through the Ntropy transaction enrichment and categorization API. Raw bank transactions go in, structured data comes out — category, merchant identity, logo, location, website, recurrence pattern, active and cancelled subscriptions, and detected income sources. A live Python quickstart shows four lines of code enriching a 1,500-row CSV (roughly 12–18 months of transaction history); the Ntropy app then surfaces the same data in a PFM-style UI. Amelia walks through Ntropy's benchmarks against GPT-4 — about 2,000× cheaper per transaction, 150× faster, and 15× more accurate on the hardest cases — before ML engineer Robin demos "Ntropy Cookie," a Discord chatbot that lets end users ask natural-language questions about their own spending: essentials vs non-essentials, savings opportunities, bank fees.
Nick Hynes
Oasis Labs demos on-chain verifiable credentials and selective disclosure on the Oasis confidential EVM. The thesis: put KYC/KYB data on-chain once with a trusted issuer (Equifax-rooted in this demo), then reuse it across web3 protocols without re-doing verification per counterparty — while giving users revocable, selective control over exactly what each verifier sees. The demo walks end-to-end through requesting a credential via identity.oasislabs.com, receiving both an NFT credential on the non-confidential Emerald chain and the full identity data stored confidentially on Sapphire (Oasis's confidential EVM, backed by trusted hardware), granting a demo website access to just the "is over 18" property, calling the isOver18 view from an Express backend, and revoking access afterwards. Broader use cases: credit score disclosure, zip-code mortgage / insurance quoting, and any flow that needs a derived fact rather than raw PII.
Raven Jiang
Arc is a B2B financing platform for software companies, pairing Arc Advance (revenue-based financing) with Arc Treasury (a deposit account for the same customer profile). This briefing walks through the Arc customer onboarding flow — NDA, API credentials for banking, accounting, and billing — and the unified data model Arc has built across all three. Live demos connect Capital One via Plaid sandbox, accounting data via Codat, and billing via Stripe, with each sync landing in Arc's GraphQL backend as canonical data events. Behind the scenes, Arc normalizes data across multiple aggregators per domain using virtual SQL views, preserves raw provider data for traceability, and optimizes connector selection for long-term stable recurring connections — because Arc re-syncs this data continuously to monitor risk on outstanding balances and surface upsell opportunities.
Jack Ryan
Jack Ryan of FinGoal demos the FinGoal Switch Kit — a wrapper around Plaid, MX Link Money, Yodlee, and other aggregators that normalizes the account and transaction shape so you can swap aggregators in minutes instead of months. The live demo starts with a Next.js app running a Plaid link against Wells Fargo, then swaps the token endpoint and frontend portal to MX Link Money — which authenticates via open banking directly in the Wells Fargo app. The migration philosophy is not a big-bang swap: FinGoal recommends leaving existing Plaid connections alone and re-linking users through the new aggregator only when a connection breaks, so net-new users and natural re-auths silently migrate without churn. The Switch Kit\'s payoff is that the success page digests the callback event string from Link Money and renders the new accounts and /transactions/get output alongside pre-existing Plaid data — same schema, same categories, same fields — with only three files changed across the entire app. The thesis: developers should not be locked into an aggregator by migration cost.
Brian Billingsley & James Armstead
Brian, co-founder of Basis Theory, walks through the Basis Theory platform — a developer-first tokenization and encryption service for any sensitive value on the internet. The demo covers tenants (test/prod, sub-customer, or geo-split workspaces), Elements (drop-in UI that keeps you out of PCI scope for cards, bank accounts, SSNs, and driver's licenses), and server-to-server apps gated by PCI Level 1 attestation. Basis Theory tokenizes any serializable value — strings, documents, PII, cards, bank accounts — and returns a consistent token shape with type, fingerprint, and non-sensitive metadata so developers can dedupe without decryption. Customers can roll their own keypair (bring-your-own-key) so Basis Theory only sees ciphertext. The reactor system ships pre-built serverless templates (Stripe, Braintree, BIN analysis via Perapt) and private reactors for custom integrations, allowing partners to tokenize a card once and route to multiple processors without ever touching the raw PAN.
Kyle Tut & Justin Hunter
Pinata manages off-chain data for blockchain applications — the infrastructure most NFT marketplaces (OpenSea, Rarible, and many others) quietly rely on. Kyle Tut, co-founder and CEO of Pinata, explains why on-chain storage is economically impractical (roughly $4M per GB on Ethereum) and how IPFS content identifiers (CIDs) give NFT creators tamper-proof references to their underlying files — you cannot silently swap out a CID the way you can a traditional Dropbox link. The demo covers Pinata's Dropbox-style upload UI, IPFS infrastructure run by Pinata to keep retrieval fast, dedicated per-customer gateways that serve IPFS content from your own domain (mypinata.cloud or your TLD) while preserving CID-level integrity, and Pinata's audit-trail approach for legitimately-updatable NFTs in gaming and interactive-art use cases.
Riddhiman Das
TripleBlind demos a cryptographic platform for privacy-preserving data and algorithm exchange — an alternative to the status-quo model of decrypting and replicating third-party data (the Capital One / Experian problem). TripleBlind\'s novel one-way encryption is homomorphic-like but roughly a trillion times faster than full homomorphic encryption, and it works across text, voice, video, and images. The data can only be used for the purpose explicitly authorized, and algorithms themselves can also be encrypted — protecting IP and training data from reverse-engineering. The live demo uses Kaggle\'s customer-transaction-prediction dataset split across three mock organizations (Standard Chartered, JPMorgan Chase, BNP Paribas) running in three different browsers. Privacy-preserving EDA lets a data scientist see shape, distribution, and plots without seeing individual records. A feed-forward neural network (dense + ReLU layers) is trained across all three parties — each granting cryptographic consent with a justification — and produces a privacy-preserving PyTorch object that can be used for local inference or secure multi-party compute predictions (both algorithm and data stay hidden from each other). Training runs in minutes where FHE would take weeks, and the code is drop-in compatible with PyTorch, Pandas, TensorFlow, and XGBoost — GDPR, CCPA, and FDIC compliant.
Peter Tapling & George Throckmorton
George (Nacha) and Peter demo Phixius — Nacha\'s business-focused data-services network for financial institutions, fintechs, and payment networks. Where consumer open banking addresses account holders, Phixius addresses the payment-provider side: exchanging data (ACH account validation, electronic payment information, etc.) between Credentialed Service Providers (CSPs) before a payment hits the rail. Phixius is point-to-point — no centralized storage, no give-to-get model. It combines standardized Nacha APIs with a blockchain/shared-ledger layer used for authentication, audit trail, and rule enforcement. Each request uses an authentication token that is effectively the address of a smart contract between the requesting CSP and responding CSP for that specific purpose. The demo runs two exchange services: an ACH account validation (routing + account → valid/invalid for Nacha), and an Electronic Payment Information exchange where Care-Free Catering at High-Rise Bank subscribes to payment info for Best Spreads at Low-Rise Credit Union — replacing the PDF-form email loop with an authoritative API exchange. A smart-contract subscription fires a notification when Best Spreads adds Zelle to their profile, proving the ongoing-update pattern.