Legacy Codebase Migration Expert

What you would do

• Design realistic migration scenarios grounded in actual enterprise systems, legacy code patterns, and technical constraints
• Analyze dependency graphs, APIs, databases, build systems, and infrastructure to map safe migration paths
• Create comprehensive testing and validation plans to detect regressions and behavioral mismatches
• Evaluate AI-generated migration code, strategies, and architectural decisions for production readiness
• Document risks, compatibility concerns, and rollout considerations for large-scale migrations

Who they want

• 5+ years of hands-on software engineering at Senior, Staff, Principal level or equivalent experience
• Direct ownership or substantial technical contribution to large-scale production migrations in mission-critical systems
• Deep familiarity with specific migration types: language upgrades (Java, Python, .NET, PHP), framework modernization, database transitions, or infrastructure changes
• Strong knowledge of testing, observability, staged rollouts, rollback planning, and how to ensure behavioral equivalence across migration phases
• Preference for experience modernizing banking, payments, trading, insurance, telecom, or healthcare systems; COBOL/mainframe expertise particularly valued

Main skills

What the interview asks about

1. 1.Assessing migration safety and risk

   You've worked on migrations where one oversight could crash a trading platform or payment system. You recognize hidden coupling and subtle incompatibilities others miss. This tests your instinct for production-grade thinking.

   For example: “A fintech firm is migrating a 15-year-old J2EE app from Java 8 to Java 17. It uses reflection and custom serialization. Outline testing priorities and describe two hidden risks a surface-level migration might miss.”

2. 2.Designing comprehensive migration scenarios

   Generic examples won't train AI well. You need to encode the messy reality of actual migrations (competing constraints, partial rollouts, state management during transitions) into your test cases.

   For example: “Design a scenario: rewrite a monolithic Django app into microservices with a new database. What technical challenges, integrations, and data consistency issues must an AI solution handle to be production-ready?”

3. 3.Spotting incomplete AI migration proposals

   AI systems might generate syntactically correct code but miss the orchestration, sequencing, and operational concerns that matter in real migrations.

   For example: “Review an AI plan for Oracle-to-PostgreSQL migration for payments. It covers schema transformation but omits dual-writes, verification, and rollback. What's missing and why does it matter?”

4. 4.Planning phased rollouts and cutover

   You've managed cutover windows in environments where downtime costs millions per minute. You understand trade-offs between safety and speed, and how to stage changes when full rollback is impossible.

   For example: “A telecom billing system moves from polyrepo to monorepo during peak hours. Describe your staging strategy, testing gates at each phase, and how you'd detect performance or correctness degradation.”

5. 5.Analyzing dependency complexity

   Legacy systems have hidden coupling, version conflicts, and compatibility layers an engineer must untangle before proposing changes. This requires reading unfamiliar code and reasoning about implicit contracts.

   For example: “Given a 20-year-old insurance system with undocumented dependencies and custom serialization, describe how you'd map what can change safely and what the AI model should verify before refactoring.”

6. 6.Creating realistic test strategies

   Testing a migration isn't just unit testing. You need regression tests, contract tests, integration tests, and sometimes shadow runs. This shows whether the model understands end-to-end validation.

   For example: “Design a testing strategy for MySQL 5.7 to PostgreSQL 14 migration in healthcare without downtime. Data accuracy is critical and compliance audits required. What test types and validation gates would you mandate?”

A task you may get

Review an AI plan for REST-to-GraphQL migration with backwards compatibility. Identify gaps in dependencies, testing, rollout, and rollback. Propose what should have been considered.

How to prepare

• Prepare a migration case study from your work: technical challenges, hidden coupling, and how you handled failures
• Review a complex legacy codebase (open source or your own prior work) and write down the dependency patterns, technical debt, and migration risks you'd highlight to an AI system
• Think through the difference between a 'working' migration (code runs) and a 'production-ready' one (monitoring, rollback, performance validated, compliance met)
• Prepare examples of subtle behavioral changes during migrations (serialization format shifts, timing assumptions, error handling changes) that testing frameworks often miss

The facts

Pay

$200/hr

Hours

Part time, 40 hours a week

Where

Remote

Field

Software Engineering

Role type

Talent network

Posted

8/22/2026

We wrote this page from the public Mercor listing. It may be out of date, so read the full posting before you apply.