AWS Platform Build and GCP Migration
Budget / Salary£3,000–5,000
TypeFreelance project
LocationRemote
Posted6 hours ago
AWS platform build and GCP to AWS migration (FastAPI, React, Bedrock, Terraform, Bitbucket)
DESCRIPTION:
We are a UK EdTech SaaS company currently running on Google Cloud. We are looking for an experienced AWS DevOps engineer to do four things: design and build our AWS platform, migrate us off GCP, move our code to Bitbucket, and modernise our release process. Phase 1 is a fixed-price project. After that, we want ongoing support and ad-hoc work from the right person.
CURRENT STACK (GCP)
- Backend: Python FastAPI containers on Cloud Run
- Frontend: React / Vite / TypeScript on Firebase Hosting
- Database: Cloud SQL PostgreSQL 15 with pgvector
- Cache: Memorystore (Redis)
- Secret Manager, Artifact Registry
- Claude models via Vertex AI (EU), plus self-hosted PII pseudonymisation services (Presidio, GLiNER)
- GitHub (3 repos) with Cloud Build triggers; Terraform partly in place
- Auth: Microsoft 365 / Entra ID SSO plus email/password
- Environments: dev, staging, demo, prod
PHASE 1 SCOPE (FIXED QUOTE)
1. AWS foundation: multi-account setup (Organizations / Control Tower), IAM Identity Center with least privilege, CloudTrail, GuardDuty, Security Hub, budgets and cost alerts. Primary region eu-west-2 (London).
2. Platform, fully in Terraform: ECS Fargate for the API and pseudonymisation services, S3 + CloudFront for the frontend, RDS PostgreSQL with pgvector, ElastiCache, Secrets Manager, ECR, WAF, private networking with VPC endpoints, CloudWatch logs, metrics and alerting.
3. Amazon Bedrock: Claude models in an EU region with IAM-scoped access, private endpoints, quotas and logging configured to our data policy. Our developers make the application code change; you provide the infrastructure and a working reference example.
4. Migration: Postgres migration with minimal downtime, DNS cutover, tested rollback plan, GCP decommission plan.
5. Source control and CI/CD: move all repos from GitHub to Bitbucket Cloud with full history. Rebuild pipelines in Bitbucket Pipelines using OIDC to AWS (no long-lived keys). Set up the Jira integration.
6. Release model: move to continuous delivery. Trunk-based development. Every merge to main is built once, tested, scanned and deployed automatically to staging. Production is a one-click approved promotion of the same artifact. Ephemeral preview environments per pull request replace dev. Demo is a stable environment running the current production release with seeded data. Zero-downtime deploys and automated rollback.
7. Backups: AWS Backup with production backups held in the UK, plus a tested restore.
8. Handover: architecture diagram, runbooks and a recorded walkthrough for our team.
IMPORTANT
- We hold ISO 27001, ISO 42001 and ISO 9001. You will follow our security policies and produce evidence for change management, access control and logging.
- We handle school data, including children's data. You must sign an NDA and data processing terms before access, and work only through accounts we provide. No production data on personal devices. UK or EU based preferred.
PHASE 2 (TO DISCUSS AFTER)
Ongoing support retainer (monitoring, patching, cost reviews, incident response) and ad-hoc projects.
TO APPLY, PLEASE INCLUDE
- Fixed price and timeline for Phase 1, split into milestones
- Your day rate for ongoing support
- Two GCP to AWS (or similar) migrations you have delivered
- AWS certifications held
- Your experience with Bedrock and Bitbucket Pipelines
Applications that do not answer these points will not be considered.
DESCRIPTION:
We are a UK EdTech SaaS company currently running on Google Cloud. We are looking for an experienced AWS DevOps engineer to do four things: design and build our AWS platform, migrate us off GCP, move our code to Bitbucket, and modernise our release process. Phase 1 is a fixed-price project. After that, we want ongoing support and ad-hoc work from the right person.
CURRENT STACK (GCP)
- Backend: Python FastAPI containers on Cloud Run
- Frontend: React / Vite / TypeScript on Firebase Hosting
- Database: Cloud SQL PostgreSQL 15 with pgvector
- Cache: Memorystore (Redis)
- Secret Manager, Artifact Registry
- Claude models via Vertex AI (EU), plus self-hosted PII pseudonymisation services (Presidio, GLiNER)
- GitHub (3 repos) with Cloud Build triggers; Terraform partly in place
- Auth: Microsoft 365 / Entra ID SSO plus email/password
- Environments: dev, staging, demo, prod
PHASE 1 SCOPE (FIXED QUOTE)
1. AWS foundation: multi-account setup (Organizations / Control Tower), IAM Identity Center with least privilege, CloudTrail, GuardDuty, Security Hub, budgets and cost alerts. Primary region eu-west-2 (London).
2. Platform, fully in Terraform: ECS Fargate for the API and pseudonymisation services, S3 + CloudFront for the frontend, RDS PostgreSQL with pgvector, ElastiCache, Secrets Manager, ECR, WAF, private networking with VPC endpoints, CloudWatch logs, metrics and alerting.
3. Amazon Bedrock: Claude models in an EU region with IAM-scoped access, private endpoints, quotas and logging configured to our data policy. Our developers make the application code change; you provide the infrastructure and a working reference example.
4. Migration: Postgres migration with minimal downtime, DNS cutover, tested rollback plan, GCP decommission plan.
5. Source control and CI/CD: move all repos from GitHub to Bitbucket Cloud with full history. Rebuild pipelines in Bitbucket Pipelines using OIDC to AWS (no long-lived keys). Set up the Jira integration.
6. Release model: move to continuous delivery. Trunk-based development. Every merge to main is built once, tested, scanned and deployed automatically to staging. Production is a one-click approved promotion of the same artifact. Ephemeral preview environments per pull request replace dev. Demo is a stable environment running the current production release with seeded data. Zero-downtime deploys and automated rollback.
7. Backups: AWS Backup with production backups held in the UK, plus a tested restore.
8. Handover: architecture diagram, runbooks and a recorded walkthrough for our team.
IMPORTANT
- We hold ISO 27001, ISO 42001 and ISO 9001. You will follow our security policies and produce evidence for change management, access control and logging.
- We handle school data, including children's data. You must sign an NDA and data processing terms before access, and work only through accounts we provide. No production data on personal devices. UK or EU based preferred.
PHASE 2 (TO DISCUSS AFTER)
Ongoing support retainer (monitoring, patching, cost reviews, incident response) and ad-hoc projects.
TO APPLY, PLEASE INCLUDE
- Fixed price and timeline for Phase 1, split into milestones
- Your day rate for ongoing support
- Two GCP to AWS (or similar) migrations you have delivered
- AWS certifications held
- Your experience with Bedrock and Bitbucket Pipelines
Applications that do not answer these points will not be considered.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.