- ▹Role: Contract Full-Stack Developer
- ▹Period: June 2026 – Present
- ▹Focus: Generative AI, full-stack architecture, application execution, and deployment
- ▹Stack: React, Next.js, TypeScript, NestJS, PostgreSQL, Prisma, OpenAI, Gemini, Docker, AWS
I work on Klego as a contract full-stack developer, leading development of an AI application builder that turns natural-language requirements into Next.js applications using autonomous coding agents. My responsibility spans the application architecture, frontend and backend implementation, model integrations, infrastructure, and deployment.
The work brings together two parts of my experience: building full-stack applications and integrating AI into a product people can use. A generated application needs somewhere to run, a preview people can inspect, and a history that makes subsequent changes understandable. Those requirements shape the engineering around the coding agents.
The product and the engineering scope
Klego starts with a natural-language description of what a user wants to build. The platform uses coding agents to turn those requirements into an application. My scope includes the systems surrounding that generation workflow: execution, previews, versioning, synchronization, authentication, and deployment.
This means working beyond a single interface or API. The frontend needs to present the application-building experience, the backend needs to coordinate the product's operations, and the infrastructure needs to support running and previewing generated applications.
Building across the stack
I work with React, Next.js, and TypeScript on the frontend, with NestJS on the backend. PostgreSQL and Prisma support the data layer. My responsibilities also include the OpenAI and Gemini API integrations used by the builder.
Owning this range of work helps keep the interfaces between components explicit. Changes in the generation workflow can affect backend state, the preview experience, and the application history. I work across those boundaries as part of the same delivery scope.
Running generated applications
I implemented Docker-isolated execution and real-time previews. These are important parts of the product because the user needs to inspect the application produced by the generation workflow.
An isolated runtime gives generated applications their own execution environment. The preview then makes the running result visible in the builder. This is the part of the system where generated code becomes something a user can examine and continue working on.
The preview and execution work connects several parts of the platform. The application code has to reach its runtime, and the running result has to be presented through the builder. My responsibility includes that connection between generated output, infrastructure, and the user interface.
Keeping track of changes
I also implemented immutable versioning and GitHub synchronization. Version history is especially relevant to an application builder because generated code can change repeatedly as the requirements evolve.
Immutable versions preserve earlier application states as the product changes. GitHub synchronization connects that work to a familiar development workflow. Together, these capabilities give the application a history that extends beyond the latest generated response.
Authentication and integrations
My work includes JWT/OAuth authentication and encrypted integrations. These sit alongside the generation and preview features as application responsibilities in their own right.
An AI builder still needs to know who is using the product and how integrations are connected. Keeping that work in the same engineering scope helps connect the AI workflow to the account and application systems around it.
Delivery through deployment
My responsibility continues through AWS deployment. I work across the architecture and delivery rather than handing off after the UI or model integration is complete.
That end-to-end scope is the main point of this project in my portfolio. It covers the coding-agent workflow as well as the application infrastructure needed to execute, preview, version, and manage the resulting work.
Current status
Klego is ongoing contract work. The scope described here includes OpenAI and Gemini integrations, Docker-isolated execution, live previews, immutable versioning, GitHub synchronization, authentication, encrypted integrations, and AWS delivery.
The project demonstrates the kind of role I can take on in an AI product team: building the model integrations while also owning the backend, interface, and deployment work around them.
Discuss a similar product
If your team is building an AI application builder or needs an engineer who can work across model integrations, application code, and deployment, get in touch.
- ▹Contact: rahul.raccog.com/contact