Five people. One pipeline.

Daniel spent years working with trading desks and market-structure research. He started on the research side of an electronic trading group, building transaction cost models and execution analytics for futures and equities, and later moved into crypto market structure: venue fragmentation, funding dynamics across perpetual exchanges and how much of a displayed order book is real.

The same problem showed up in every engagement. Before any research could start, someone had to collect, repair and align the data, and the repairs were rarely documented. He founded Finance Data Cloud in 2025 to do that work once, with capture timestamps on every message and gaps recorded instead of filled.

He runs the research program behind our extended fields, including the out-of-sample study on Hyperliquid perpetuals, and takes the first call with most pre-registered partners. If you want to argue about how VPIN should be bucketed, he is the right person to write to.

Chris has built data platforms for fintech and exchange infrastructure teams. He led the data platform group at a London payments company, where his team ran the reference-data and reconciliation pipelines behind several hundred million transactions a year. He then spent four years as product lead for market data distribution at an exchange technology vendor, shipping feed handlers and historical data products to brokers and trading firms.

That second role taught him how data contracts work from the venue side: what exchanges allow vendors to redistribute, how licensing differs by dataset and why the fine print matters to a compliance team. He joined Daniel as co-founder in 2025.

At Finance Data Cloud he owns the product and everything commercial: pricing, the pre-registered partner program, venue relationships and the data license. He also writes most of the onboarding material that early access partners will receive in Q1 2027.

Maya owns the normalization layer: the schema that makes a trade from any two venues look the same. Before Finance Data Cloud she built the tick-data warehouse at a European energy trading firm, then spent three years at a crypto analytics company normalizing exchange and on-chain data for institutional clients.

She wrote schema v0.1 in March 2026 and the v0.9 draft that v1 will be cut from. The decisions with her name on them include decimal128 for every price and size, one symbology across spot, perpetuals, futures and options, and keeping the exchange's native symbol in venue_symbol so nothing is lost in translation.

Each new venue starts with Maya reading its API documentation end to end and writing down every field that does not map cleanly. Those notes become the venue adapter, the data-quality checks and the per-venue page in our docs.

Lukas builds the APIs and streaming services, with a focus on low-latency delivery and gap-free replay. He spent six years on backend systems at a Berlin brokerage and then a market data vendor, where he built order-routing gateways and a WebSocket fan-out service that pushed quotes to tens of thousands of client connections.

He joined in January 2026 and owns everything between our normalized store and your code: the REST API, the WebSocket stream, bulk S3 delivery and the Python SDK. The live stream target of under 5 ms added latency (median, same region) is his number to hit.

His current focus is replay. A request with live=True backfills from the archive and hands over to the live stream without a duplicated or missing message at the seam, and any known gap in the requested window comes back with the data.

Priya runs capture infrastructure across regions so no message is missed, from collectors to long-term storage. She previously ran infrastructure for a digital asset exchange in Asia, covering its matching engine hosts, market data gateways and on-call rotation, and before that worked as a site reliability engineer on a large consumer platform's data team.

She brought up our Hyperliquid collectors in Tokyo and Ireland in November 2025 and has kept them recording since. She also set up the PTP-disciplined clocks behind every ts_capture value, the merge that reconciles the two regions by venue sequence number, and the storage tiers that keep raw messages untouched next to the normalized tables.

Next on her list is the N. Virginia (us-east-1) region planned for Phase 2, and the capture design for the venues that follow Hyperliquid once data agreements with those exchanges are in place.