Data & Migration Engineer

Apply Description Planet Depos ships enterprise software for the legal industry, and the certified record is our product: if the data is wrong, the business is wrong. This role owns the data layer of our platform, and that is three jobs in one. Design: the data models and data contracts that our product teams, our low-code estate, our integrations, and our analytics all build against. Migration: we are replacing the legacy system the business runs on, which means moving two decades of operational data out of an undocumented database safely and provably while the business keeps running on it. Quality: the feeds, reconciliation, and monitoring that make the data trustworthy every day, not just on cutover day. ESSENTIAL JOB RESPONSIBILITIES: Own the target data model in partnership with the Group Architect: the canonical entities the platform is built around, identity and access-bearing structures, and the boundaries with systems that own their own truth (CRM, finance) Own the data workstream of the legacy-platform replacement end to end: profiling, mapping, extraction, transformation, load, verification, and a staged, rehearsed cutover Build and operate the ongoing synchronization that keeps the new platform current through the transition, engineered for a fragile legacy source: rate limits respected, timeouts and retries handled, silent feed breaks made impossible by monitoring Design the data contracts our integrations run on: the supported surfaces and interchange formats that let customer systems and partner tools consume our data without bespoke plumbing Give the low-code estate a safe foundation: supported data surfaces for the citizen-developer applications and flows the business runs on, replacing direct database access over time Build the reconciliation and data-quality evidence: automated counts and checksums, exception queues, and human-readable verification reports Create and maintain sanitized and synthetic datasets from production, and manage dataset promotion across environments, so non-production work never requires production data Lay the analytics foundation with the platform team: safe read paths (replicas, sanitized exports) and data structures the business can query without touching production Steward database design standards with the platform team: DevOps owns infrastructure operations; this seat owns how data is modeled, accessed, and kept trustworthy Document the models, mappings, and pipelines so they are auditable and transferable, readable by both humans and agents REQUIRED WORK OPERATING STANDARDS: Evidence over assertion. The number is right because the reconciliation proves it, not because the script ran. You make your work legible: entity models, mapping documents, runbooks, and pipelines someone else could operate. You raise solutions alongside your objections, with an owner and a date. You have worked in a fast-moving environment and stayed effective when priorities shifted under you. You can disagree, and then commit wholeheartedly once a call is made. You treat access to sensitive data as a responsibility to design around, not a convenience to accumulate. REQUIRED EXPERIENCES: You have moved production data between live systems without losing any. You have executed at least one legacy-to-modern migration of a business-critical system, with the discipline that implies: reverse-engineering an undocumented source schema, building the field-level crosswalk from what the data actually contains, staged runs, parallel verification windows, documented reconciliation, and a rollback plan you could actually invoke. You know a migration is done when the evidence says so, not when the script exits zero. Evidence looks like: "Migrated the operational database of a revenue-carrying system with row-level reconciliation and zero unexplained variance," or the mapping document for a several-hundred-table legacy schema that became the build contract for the replacement. In an interview, you can walk through the migration that went sideways, what the reconciliation caught, and what you changed. You have designed data models and data contracts that a whole estate runs on. You have owned an operational data model with many consumers: product applications, low-code and citizen-developer workloads, third-party integrations, and analytics, each pulling in a different direction. You know how to define the canonical entities, decide which system is the source of truth for what, and draw the boundary where another system's model stops being yours. You design the supported surfaces (views, APIs, contracts) that let those consumers build safely instead of reaching into tables directly, and your models outlive the applications that first used them. Evidence looks like: "Designed the core entity model a company's next three products were built on," or "replaced direct database access for a sprawl of internal tools with governed, supported data surfaces without breaking the business." In an interview, you can defend a source-of-truth boundary you drew and describe the consumer you said no to, and why. You treat data quality as an engineering discipline, with evidence. Verification is your deliverable, not an afterthought: automated reconciliation (counts, checksums, exception queues), drift monitoring on ongoing feeds so a silent break is impossible, and verification reports a non-engineer can read and trust. The same discipline extends past the migration: the daily synchronization, the analytics foundation, and the data the business reads every morning. You handle confidential data like it matters, because here it does: litigation material under client data-protection commitments, least-privilege access, sanitized data everywhere outside production. And you have built that yourself: created sanitized or synthetic datasets from production systems and managed which datasets live in which environment, so non-production work never needs production data. Evidence looks like: "Built the reconciliation harness that ran nightly during a nine-month parallel run and caught every variance before the business did," or "built the sanitized-dataset pipeline that let every non-production environment drop real customer data." COMPENSATION RANGE: $150,000 – $205,000 annual base compensation, commensurate with experience. Benefits include Medical, Dental, and Vision coverage; Life insurance (Voluntary Term and Whole Life); Voluntary Long Term Disability; paid time off and paid holidays; 401(k); Employee Assistance Program (EAP); and Maternity Leave. Requirements 5+ years in data engineering, database engineering, or migration-heavy roles Expert SQL (SQL Server and Postgres are both in play here), plus a scripting language for pipeline work (Python or TypeScript) Demonstrated operational data modeling: you have designed schemas and data contracts that multiple teams or systems consumed, and can speak to the tradeoffs you made At least one completed production migration of a business-critical system, with reconciliation you can speak to in detail Experience creating sanitized or synthetic test data from production systems and managing datasets across environments Experience running extraction or synchronization against fragile or rate-limited legacy sources Experience handling regulated or confidential data; legal-industry experience is a plus, not a requirement US-based, as a hard requirement of our client data-protection commitments Salary Description $150,000 - $205,000 annual base compensation