Data Solutions Architect

Description

The bar is high – bring a curious and forward-thinking mindset and we'll give you the platform to define what comes next.

The perks

• Matched giving for your fundraising activity

• Flexible working hours and work-from-home opportunities

• Performance-related bonuses

• Insurance and medical plans

• Career-focused technical and leadership training in class and online, incl. unlimited access to LinkedIn Learning platform

• Contribution to gym memberships and more

• Free Lunch/Snacks

• A day off on your birthday

• Two days’ volunteering leave per year.

Person specification

• A collaborator who works across IT and business teams to realise integration design and evolve the strategy jointly rather than in isolation.

• Sound, pragmatic judgement — balancing pace, risk and commercial value to reach decisions that are well informed, rigorous and actionable.

• Builds credibility and influence with stakeholders through technical substance and clear communication.

• A good facilitator — draws out problems, builds consensus, and secures commitment from the people who have to deliver.

• Curious and forward-thinking, with the appetite to grow scope over time.

What you’ll do

• Own the solution architecture and high-level design for Post Trade regulatory reporting data — from trade and lifecycle event capture on the enterprise message bus, through the lakehouse layers, to submission, acknowledgement handling and reconciliation against trade repositories.

• Design and deliver solution architecture for assigned domains on the GCP lakehouse — from event ingestion off Kafka, through Bronze/Silver/Gold, to consumption by analytics, operational and regulatory reporting workloads.

• Model the data. Dimensional and canonical modelling, conformed dimensions, slowly changing dimensions, grain determination, and the trade-off between nested and repeated structures versus normalised models in BigQuery. Design to the enterprise canonical standard — extending it through the governance forum where a domain genuinely needs something it does not cover.

• Own schema contracts in practice. Avro schema design and review; compatibility mode selection (backward, forward, full and their transitive variants) and the producer/consumer upgrade ordering each one implies; union and default-value design for evolvability; topic, key, partition and compaction design; and the operational consequences of getting any of those wrong.

• Design ingestion for both batch and streaming. Watermarks, windowing, triggers and allowed lateness for out-of-order arrival; replay and restatement; idempotency and deduplication strategy; dead-letter and poison-message handling; and honest reasoning about where exactly-once guarantees actually stop — they do not extend past the Kafka boundary to an external sink.

• Build data quality and profiling into the design — DQ gates at layer boundaries, reconciliation and break detection, completeness and accuracy controls, and the lineage and audit evidence that regulated consumers, internal audit and regulators will ask for.

• Apply the governance model. PII classification, BigQuery policy tags, column-level and row-level security, dynamic data masking, IAM/RBAC design, CMEK and key management — including the design tension where privacy obligations and multi-year regulatory retention pull in opposite directions.

• Work with infrastructure and platform engineering to specify and guide the setup of the runtimes the data architecture requires.

• Produce the artefacts and take them through review. HLDs, interface specifications, ADRs and reference designs, authored to a standard that survives formal architecture governance and your own absence from the room.

• Work across teams — with producing service teams on event contracts, with data and database engineering, messaging and data science, and with business and compliance stakeholders on what the data has to mean.

• Contribute to the architecture - design reviews, standards feedback, knowledge sharing, and mentoring engineers into architectural thinking.

What you’ll need for this role

• 10+ years in data engineering and data architecture, with hands-on delivery of cloud data solutions from design through to production, and demonstrable ownership of at least one end-to-end data solution in a regulated environment.

• Post Trade or regulatory reporting data experience. Direct delivery experience on at least one of MiFIR, EMIR (including Refit), ASIC, MAS, SFTR or CFTC reporting. Reportable-event determination, lifecycle and correction semantics, identifier management (UTI, UPI, LEI, ISIN), pairing and matching, reconciliation against trade repository acknowledgements and rejections, back-reporting, and a working grasp of what a completeness or accuracy breach actually costs to remediate.

• Deep, hands-on GCP. BigQuery is the centre of gravity: partitioning and clustering strategy, slot reservations and cost governance, authorised views and datasets, materialised views, table snapshots and time travel, the Storage Write API, policy tags, column-level and row-level security, dynamic data masking, and CMEK. Plus Cloud Composer/Airflow (idempotent task design, backfill and catch-up semantics), Dataflow or Dataproc, Datastream for change data capture, GCS, Cloud Run and Cloud Functions, Pub/Sub, IAM, VPC Service Controls, and Dataplex for classification, data quality and lineage. GCP depth is what we are hiring for; breadth on other clouds is welcome but does not substitute for it.

• Medallion lakehouse design in production — Bronze/Silver/Gold with immutability at Bronze, progressive quality enforcement, deterministic reprocessing, and a defensible position on where transformation logic is allowed to live.

• Kafka at enterprise scale with a governed schema registry (Cloudera, Confluent or equivalent) — Avro schema design and evolution, compatibility modes and their operational consequences, event-driven ingestion into a cloud data platform, consumer group and lag management, compacted topics and tombstone semantics, and the ordering and delivery guarantees Kafka does and does not provide.

• Data governance in a regulated environment — IAM/RBAC frameworks, PII classification and encryption, data retention design, audit logging, and compliance-driven design under FCA, GDPR or equivalent supervision.

• Architecture artefacts that have passed formal review — HLDs, ADRs, interface specifications, reference designs. You can evaluate competing options, articulate trade-offs without hedging, decide under ambiguity, and write the decision down.

• Cloud integration patterns — REST API, MFT, event streaming, cross-cloud AWS↔GCP — including Workload Identity Federation, VPC perimeters, and JWT-based service-to-service authentication.

• Working command of core data platform concepts — data lake, ODS, data warehouse and lakehouse — and a clear view of when each is the right answer rather than a default.

• Stateful stream processing with Apache Flink or Spark Structured Streaming — state stores, checkpointing, and event-time processing at scale.

• Enough Java and event-driven / microservices fluency to review service designs and challenge engineering teams credibly. This is not a hands-on development role.

• Oracle RDBMS, PostgreSQL and MongoDB, in the context of a legacy estate still carrying business-critical data.

• dbt, Terraform or equivalent infrastructure-as-code.