Full-Stack Engineering & DevOps for a National Bank's Workforce Platform

Overview

Working inside a regulated financial institution is categorically different from working in a startup. Every infrastructure change passes through a change advisory process. Access to production is time-limited and audited. The risk appetite is asymmetric — the cost of a mis-deploy is far higher than the cost of moving slowly.

At the same time, the engineering teams inside large organisations face the same fundamental problems as everyone else: pipelines that are not as reliable as they need to be, infrastructure that accumulated without a clear owner, and pressure to ship product faster than the governance processes comfortably allow.

This is an ongoing embedded engagement — senior full-stack and DevOps capacity working inside the delivery team on the bank's workforce re-engagement and learning platform.

What the Platform Does

The platform manages structured workforce re-engagement and career development: employees on extended leave, undergoing rehabilitation, or pursuing internal career transitions are tracked through milestones, learning content and support interventions. HR managers, occupational health practitioners, line managers and the employees themselves each interact with different surfaces of the system — a React web application, a React Native mobile app for the learning experience, and a Django backend orchestrating workflow (including Flowable-based process automation).

The infrastructure runs on AWS — Amplify for the web frontends, RDS, Lambda, DynamoDB, ECS and S3, with AppSync and API Gateway at the integration edge. The stack is not unusual; the operational requirements are.

What We Do on the Engagement

Full-stack feature delivery

Day-to-day engineering across all three surfaces — React web, React Native mobile and the Django backend — within agile sprint ceremonies and weekly releases to UAT and production. Feature work, bug resolution and client-facing iteration happen inside the bank's delivery process, not around it.

Pipeline Reliability and Release Governance

In a regulated environment, the CI/CD pipeline is not just an automation tool — it is part of the audit record. Which build ran, who approved it, what was in the diff, and when it hit production are questions that compliance teams ask retrospectively.

Our work here covers:

• Pipeline hygiene: ensuring build outputs are deterministic and traceable from commit to deployed artifact.
• Environment promotion gates: staging → UAT → production promotion paths with explicit approval steps, not implicit "merge to main" triggers.
• Rollback readiness: every deployment should have a tested rollback path. We review changes for reversibility before they go to the CAB (change advisory board).

Cloud Infrastructure and Cost

IAM sprawl, unused resources and configuration entropy accumulate faster in teams that are shipping under pressure. Part of the role is periodic review of:

• IAM role and policy least-privilege alignment
• AWS cost analysis — Cost Explorer reviews, right-sizing and eliminating idle spend
• RDS configuration and backup posture against the DR policy
• Lambda, ECS and Amplify configuration across environments

In a banking context, findings from these reviews feed into the internal risk register rather than a simple backlog ticket — the output format matters as much as the finding.

Working Model

Delivery happens inside the bank's process:

• Sprint ceremonies, client meetings and weekly UAT/production releases
• Async review of PRs and architecture decisions
• Written findings for anything that needs to go into the risk or compliance record

We do not hold production access as a standing permission — access is requested per change window, granted, used, and revoked. This is the right model for a banking environment and we operate accordingly.

• React
• React Native
• Django
• AWS Amplify
• Lambda / ECS / DynamoDB
• AppSync / API Gateway
• CI/CD
• Change Governance

Why This Kind of Engagement Works

Large organisations often cannot hire senior engineers quickly — the procurement process alone takes months. An embedded retainer lets the internal team access senior full-stack and DevOps capacity immediately, on-demand, without a headcount commitment. Knowledge transfer happens naturally as part of the work, building internal capability rather than dependency.

The engagement has been running for well over a year across web, mobile and backend — handling the day-to-day delivery alongside the harder questions: the periodic reviews, the cost work, and the changes that need a senior second opinion before they go to the CAB.