Streaming Data Hackathon For Financial Services

about this event

One day in London, the biggest Financial Hub in Europe. Solve real banking problems with streaming data: fraud interception, transaction enrichment, mainframe offloading, customer 360. Guided by the engineers who created Apache Flink.

One day. Five real challenges drawn from real financial services workloads. Ververica engineers at the whiteboard and in the code with you.

You will not sit through a platform pitch. You will pick a problem your institution actually has, design and build a solution against it in a running Ververica Platform environment, and present your architecture and trade-offs at the end of the day.

Real-time data can grow revenue, cut operational costs, and protect the business. That is a claim. Spend a day and test it.

Who should be in the room

Developers. Solution architects. Data architects. Platform engineers. Innovation and modernization teams, finally together.

Banks, insurers, fintechs, and financial platforms.

Send a cross-functional team. This works best when application, infrastructure, and architecture people from the same institution work the same problem together. One day of shared context beats six weeks of forwarded slide decks. Teams of three to five.

Coming alone is fine. We will match you at 09:30.

Prerequisites

You need working SQL. That is the real bar.

You do not need prior Apache Flink experience. Challenges are built to be solved in Flink SQL, and a guided lab runs in parallel all day for anyone who wants to learn the engine rather than compete on it.

If you write Java, Scala, or Python, bring it. You will have the DataStream API and PyFlink available.

Five Financial Service Challenges

1. Fraud interception Score a live transaction stream or build a Feature Store. Detect suspicious patterns, enrich in-flight, and raise an alert before authorization completes. The production bar in this space is under 10ms. See how close you get.

2. Transaction enrichment and classification Normalize messy payment descriptions. Classify events. Enrich from reference data and push clean output downstream. Solvable end to end in Flink SQL.

3. Real-time AI Agents Automate decisions leveraging Real-time AI Agents for transactions and customer care, integrating VectorDB and LLM services in a streaming pipeline.

4. Customer 360 in motion Join event streams to reference data and maintain a customer profile that is correct at read time, not correct as of last night.

5. Compliance and audit monitoring Track critical business events, delivery guarantees, and exception conditions continuously. DORA, Basel, and MiFID II reporting on data that has not gone stale.

Running alongside: the Streaming SQL Lab Not a competition track. A guided path through Flink SQL and CDC with an engineer, for teams who want the fundamentals before they build.

Two ways to submit Both are judged. Both are legitimate. Build. A working pipeline on the provided environment. Design. An architecture, with the trade-offs written down and defended. For teams whose real constraint is a decision, not an implementation. Architects should not feel obliged to out-type a graduate developer to be taken seriously. Judge the thinking.

Your data stays where it is You will not bring customer data, and we will not ask for it. Every track ships with a realistic synthetic dataset: payment flows, CDC streams off a mock core, reference tables, and injected anomalies. Nothing you produce leaves the sandbox unless you choose to take it with you. The environment runs in the EU. Engineered in Europe. Sovereign by design.

What we provide A provisioned Ververica Platform environment, live before you arrive. Synthetic financial datasets and running generators for all five tracks. No time lost building fixtures. Starter repos for Flink SQL, PyFlink, and the DataStream API. Ververica engineers in the room all day. Not as sales. As mentors on SQL, connectors, state, performance, and operations. Mentor checkpoints at 11:00, 13:30, and 15:30. Every team, no sign-up. Food, coffee, and drinks after the presentations.

Schedule Opening session: set business context, explain the challenge themes, and align on judging criteria Technical kickoff: short enablement on architecture, tooling, datasets, and success patterns Team hack time: groups build a prototype, architecture, or working pipeline Lunch Time Mentor checkpoints: Ververica experts support teams on design, SQL, connectors, state, performance, and operations Final presentations: each team presents the solution, trade-offs, and business impact Wrap-up: summarize patterns observed and identify next-step use cases for follow-up

Drinks and Networking

What every team leaves with A written architecture review of your solution from a Ververica engineer, delivered within five working days. Ververica Academy certification vouchers for every participant. Platform credits and a sandbox that stays live for 30 days so the work does not die at 18:00. Your track's reference solution and dataset, to reuse internally. 🏆 PRIZE for the winning team!

Venue 📍Location: London 🏟️Venue: The exact address will be shared closer to the event date. ⚠️ Approval required. Registration closes when we reach the cap. Free to attend.

tracks

schedule

tickets

prizes