> Markdown version of [/jobs/ext/4653567-ai-data-modeler-w2](https://www.wearedevelopers.com/jobs/ext/4653567-ai-data-modeler-w2). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # AI Data Modeler (W2) - **Company:** The Bridge Ltd - **Location:** Malvern, PA, United States - **Experience:** Expert - **Contract:** Contract - **Skills:** AWS Textract, Microsoft Excel, Artificial Intelligence, Amazon Web Services, Identity and Access Management, Python (Programming Language), Regression Testing, Software Engineering, Management of Software Versions, Data Logging, Large Language Models, Prompt Engineering, Agentic-AI, Amazon Bedrock, Evaluation of Large Language Models, Cloudwatch, Amazon Simple Queue Service (SQS), Web Api - **Published:** November 10, 2026 - **Apply:** https://www.dice.com/job-detail/7b984b30-543a-49dc-aedb-12e58e0c160b ## About the Role 7+ years of software engineering experience with Python as your primary language * Hands-on experience with Amazon Bedrock - model invocation, Agents Knowledge Bases, and Guardrails * Experience with prompt engineering - designing, versioning, and evaluating prompts in a production context, not iust experimentation * Familiarity with LLM evaluation patterns: ground truth datasets, accuracy measurement, regression testing for model outputs Strong AWS fundamentals: Lambda, $3, SQS, SNS, EyentBridge, CloudWatch, IAM Secrets Manager * Experience with agentic patterns and tool use - building systems where the model calls external APIs as part of a reasoning chain * Comfortable working with structured and unstructured document formats (PDF Excel, DOCX) and document intelligence pipelines (AWS Textract or equivalent) * AWS ML certification is a plus; a demonstrated track record working with Bedrock and Foundation Models in production carries equal weight * Understanding of how to build observable Al systems - logging model inputs/outputs, tracking token costs, detecting accuracy drift in production