Skip to content
Nubent

/Apps and development on AWS

AWS application development with Terraform and CI/CD

We design and build cloud-native software on AWS: backends, APIs, frontends, databases and the pipeline that ships them to production. The same team writes the application code and the Terraform, so your product is operable, secure and cost-transparent the moment it goes live.

/Diagnosis

Signs it is time

Your roadmap outgrew your team

The product is validated and customers are waiting, but your in-house team cannot ship new features and keep the infrastructure running at the same time.

The code works, the cloud does not

The app runs fine locally, but every deploy is manual, environments drift apart and nobody is quite sure which AWS resources support what.

You are starting a new product

You are about to launch a SaaS, a public API or an internal app and want to choose correctly between ECS, Lambda and EKS before writing the first line.

Tech debt is slowing releases

A monolith that is hard to deploy, with no tests or observability, makes every change risky and pushes your time to market further out every quarter.

/What we deliver

Concrete deliverables, in your repository

  • Application architecture

    Architecture decision records (ADRs) covering compute, data, integration and security, plus diagrams and a cost estimate built with the AWS Pricing Calculator.

  • Backend and APIs in production

    Services on ECS Fargate or Lambda, REST or GraphQL APIs with API Gateway or AppSync, authentication with Amazon Cognito and versioned OpenAPI contracts.

  • Globally distributed frontend

    Next.js or React web apps served from S3 and CloudFront, or with AWS Amplify Hosting, with domains in Route 53 and ACM certificates.

  • A deliberately designed data layer

    Aurora PostgreSQL, DynamoDB or both depending on the access pattern, with backups, KMS encryption, replicas and automated schema migrations.

  • CI/CD and infrastructure as code

    Pipelines on GitHub Actions or CodePipeline, identical dev, staging and production environments in Terraform, and zero-downtime deployments.

  • Operations ready at launch

    Logs, metrics and traces in CloudWatch and X-Ray, actionable alarms, runbooks and a documented handoff so your team can maintain it.

/How we work

Our process

  1. Technical discovery

    We map the product, its users, non-functional requirements and budget constraints. If you already have an AWS account, we review it with read-only access.

  2. Architecture and delivery plan

    We define the architecture, the data model and an incremental roadmap, aiming to get a first piece into production as early as possible.

  3. Built through pull requests

    Application and infrastructure code in your repository, reviewed through pull requests, with automated tests and continuous deployment to every environment.

  4. Launch and operations

    Load testing, a go-live checklist, active monitoring during launch, and then ongoing support or a handover to your team.

Software development on AWS: code and infrastructure together

Most development projects split the people who write the application from the people who deploy it. The result is familiar: software that works locally and breaks in the cloud, environments that do not match, and AWS bills nobody can explain. At Nubent the same team designs the code and the infrastructure, because on AWS the two cannot be separated.

That means every design decision accounts for how it scales, what it costs and how it is observed. Choosing between an Amazon SQS queue and a synchronous call, or between DynamoDB and Aurora, is not a developer preference: it changes your product's latency, resilience and monthly bill.

Everything lives in your repository. The application, the Terraform modules, the pipelines and the documentation are under your control from the first commit, reviewed through pull requests. If you decide to bring development in-house, your team inherits a system it can understand, not a black box.

How we choose compute: containers, serverless or Kubernetes

There is no single answer. For most web apps and APIs with steady traffic, Amazon ECS on Fargate offers the best balance of operational simplicity and control. For spiky workloads, asynchronous jobs or event-driven integrations, AWS Lambda removes servers entirely and bills per invocation.

Amazon EKS makes sense when you already run Kubernetes, need portability, or have many teams deploying dozens of services. Proposing it for a product with three services usually adds complexity with no payoff. We record that choice in an ADR, with estimated costs and the alternatives we ruled out, so you can revisit it when the context changes.

  • ECS Fargate: APIs and web applications with sustained traffic.
  • Lambda: events, scheduled jobs, unpredictable spikes and lightweight backends.
  • EKS: platforms with many services, many teams or portability requirements.

Frontends on CloudFront and Amplify, APIs and databases

On the frontend we use Next.js or React distributed through Amazon CloudFront, either with an S3 bucket and edge functions or with AWS Amplify Hosting when the team values fast iteration and per-branch previews. We add AWS WAF, security headers and well-tuned caching so the experience is fast wherever your users are.

APIs are designed contract-first: OpenAPI for REST on API Gateway, or GraphQL schemas on AppSync. Authentication lives in Amazon Cognito or your own identity provider. For data, we choose by access pattern: Aurora PostgreSQL for relational models and reporting, DynamoDB for key-based access at scale, and ElastiCache when latency demands it.

CI/CD pipelines and production releases you can roll back

No project is done until a pipeline takes every change to production automatically and reversibly. We set up GitHub Actions or AWS CodePipeline with tests, static analysis, dependency scanning and blue/green or canary deployments depending on the service. Environments are created with Terraform and are identical to one another.

Before launch we run load tests, validate alarms and write runbooks for the most likely scenarios. On go-live day we are watching the dashboards alongside your team. Afterwards you can keep working with us on an ongoing basis or receive a documented handover, pairing sessions included.

/FAQ

AWS Application Development: frequently asked questions

Another question? Write to hello@nubent.com and an engineer will answer.

How much does it cost to build an application on AWS?

It depends on functional scope, integrations and availability requirements. That is why we start with a short technical discovery that ends in an architecture, an incremental roadmap and an estimate of both effort and monthly AWS cost. You decide with real numbers before committing the full budget, and you can prioritize what to build first.

Which is better for a new application: ECS, Lambda or EKS?

For most new web apps and APIs, ECS on Fargate is the simplest and most predictable starting point. Lambda fits event-driven or highly variable workloads. EKS is justified when there are many services and teams, or when Kubernetes expertise already exists. We decide based on your actual requirements and document the decision.

Do the code and infrastructure belong to my company?

Yes. We work in your repositories and your AWS accounts from the start. The application, Terraform modules, pipelines and documentation are yours. We do not use proprietary platforms or closed frameworks, so you can continue development with your in-house team or another provider without rewriting anything.

Can you join an existing development team?

Yes, it is one of the most common scenarios. We plug into your rituals, your repository and your ticketing tool, take on complete features or the infrastructure and CI/CD work, and leave behind practices your team adopts: pull request reviews, reproducible environments, observability and automated deployments.

What is the difference between AWS Amplify and CloudFront with S3?

Amplify Hosting handles builds, deployments, per-branch previews and SSR in one integrated service, ideal for frontend teams that iterate quickly. S3 with CloudFront gives more control over caching, edge functions, WAF and costs at scale. We choose based on expected traffic, the framework and how much operational control your team needs.

Next step

Let's talk about your AWS Application Development project

30 minutes with an engineer. Concrete next steps, whether or not you work with us.