AI Engineer - AI Model Integration Build (107-08SENG-01)

Cloudary is a cloud consulting firm founded and led by senior cloud architects. We embed Principal-Vetted engineers directly into high-performing tech teams — acting as agile, hands-on experts driving critical data and AI initiatives.

We are looking for an AI Engineer to deliver a scoped AWS model-integration project for a multi-cloud AI filmmaking platform. This is a proof-of-concept engagement, model-integration-as-a-service: a thin AWS Lambda + API Gateway proxy layer that plugs into the client's existing Firebase/GCP frontend, without migrating their backend or disrupting their release process. You bridge the gap between theoretical data science and practical software engineering, turning AI concepts into a real, working integration under a tight 3–4 week window.

What You Will Do

• Build the AWS Proxy Layer: Architect and deploy a serverless proxy (AWS Lambda fronted by API Gateway) provisioned via Infrastructure-as-Code (AWS CDK or CloudFormation), mirroring the client's existing model-call pattern with no changes to their Firebase deployment pipeline.

• Add 9:16 Vertical Parameter Support: Extend six model integrations (Veo, Sora, SeedDance for video; GPT-Image, Nano Banana, FLUX for image) to support vertical aspect ratio, passed through to each provider's API with no changes to existing prompt/style/camera controls.

• Integrate SeedDance 2.5: Build a direct ByteDance/ModelArk REST API integration (bearer-token auth), wrapped in a dedicated Lambda function and surfaced in the client's UI using their existing dropdown pattern.

• Implement Content Moderation Pre-Check: Add an Amazon Bedrock-based moderation step ahead of FLUX and SeedDance calls specifically, screening for celebrity/political likeness and graphic content (final mechanism — Guardrails policy, LLM classification, or combination — defined during design).

• Validate End-to-End: Exercise the full flow through a new demo project — story ideation → storyboard image generation → 9:16 SeedDance video generation → MP4 export — validated in a Beta environment.

• Document Decisions: Confirm and document technical assumptions (e.g., no-train/no-retain flag verification for the ModelArk API) so deliverables are defensible and auditable.

What We Are Looking For

• Production AWS Experience: Hands-on with Lambda, API Gateway, and IAM/networking in real deployments — not just certification.

• Infrastructure as Code: Comfortable provisioning environments via AWS CDK or CloudFormation.

• Amazon Bedrock Familiarity: Experience with Bedrock Guardrails or LLM-based classification for content moderation use cases, or the ability to ramp quickly.

• REST API Integration Skills: Proven ability to wrap third-party/vendor APIs (bearer-token auth, REST) into clean, production-ready service layers.

• Full-Stack Awareness: Comfortable working alongside an existing React/Firebase frontend without owning frontend design or development work.

• Delivery-Oriented Mindset: Able to work independently within a short, fixed-scope engagement (15–20 business days / 20–25 calendar days) and defend technical trade-offs directly with the client.

• Working English: For direct client communication and technical validation sessions.