RTM Developer

RTM Knowledge Engine AWS and Software Engineering Resource Profile

Internal resource description for the proof of concept

Core focus | AWS and Bedrock | Python and APIs | Knowledge retrieval | Structured application analysis

Role purpose

Provide experienced hands-on engineering capacity to build the first RTM Knowledge Engine proof of concept. The work will make approved RTM documentation searchable, create structured views of selected application artifacts, and establish an initial map of relationships among overlays, procedures, and data files. The resource will work with RTM specialists who confirm domain meaning and identify available compiler and interpreter evidence.

Initial scope

The POC will use an approved sample of approximately 1020 representative overlays, their associated file definitions, and selected RTM documentation. It will focus on repeatable ingestion, source traceability, structured facts, and a small set of deterministic queries. The resource is not expected to independently reverse engineer or modify RTM compiler or interpreter components.

Responsibilities

• Design and implement a controlled AWS development environment using services such as IAM, Amazon S3, AWS Lambda, Amazon API Gateway, and Amazon CloudWatch. Apply approved access, data handling, logging, and cost controls.
• Configure Amazon Bedrock Knowledge Bases for approved RTM documents. Prepare metadata, test parsing and chunking, and evaluate retrieval against known-answer questions with usable source citations.
• Develop maintainable Python scripts and small APIs to ingest and normalize approved compiler or interpreter artifacts, such as listings, cross-references, intermediate representations, and diagnostics.
• Work with RTM SMEs and compiler specialists to determine which artifacts are available, what facts they can establish, and where evidence is incomplete. Record assumptions and unknowns.
• Define a versioned data model for selected RTM entities and relationships, including overlays, procedures, files, clusters, segments, fields, calls, and references. Retain source version and location, extraction method, confidence, and SME review status.
• Implement deterministic queries for selected questions, such as which files or procedures an overlay uses and where a field is referenced. Validate outputs against SME-confirmed examples.
• Once retrieval and structured queries can be tested independently, integrate them into a limited agent workflow using Amazon Bedrock AgentCore or Strands where appropriate. Ensure explanations distinguish evidence from inference.
• Maintain source control, tests, technical notes, evaluation results, known limitations, and repeatable run and handover instructions.

Required capabilities

• Demonstrated AWS engineering experience, including IAM, S3, and hands-on use of services such as Lambda, API Gateway, and CloudWatch in a controlled development environment.
• Practical experience building Python applications or data-processing scripts and integrating REST APIs; able to work with JSON and versioned schemas.
• Hands-on knowledge of Amazon Bedrock Knowledge Bases or comparable retrieval augmented generation solutions, including document ingestion, metadata, chunking, retrieval testing, and source grounding.
• Experience designing structured data models and repeatable data pipelines, with attention to provenance, source locations, versioning, and validation.
• Ability to build or support deterministic services and queries that can be checked against source evidence or SME-confirmed examples.
• Strong technical communication and collaboration skills; can explain design choices, surface risks, and document gaps without treating assumptions as facts.
• A disciplined approach to access control, security, approved data handling, testing, and cost awareness.

Useful supporting experience

• Graph modelling, relational or document databases, and dependency or data-lineage queries; familiarity with Amazon Neptune is useful if graph traversal is justified.
• Compiler or interpreter outputs, static analysis, parsers, abstract syntax trees, or legacy application environments.
• Amazon Bedrock AgentCore, Strands, model tool use, or agent evaluation.
• A lightweight front end using React, Amplify, or Cognito if a user interface is included in the approved POC scope.

Staffing approach

One experienced resource may cover the work if they combine AWS and Bedrock retrieval skills with strong Python and API engineering. If that combination is not available in one person, split the assignment between an AWS Knowledge and RAG Engineer and a Software Integration and Structured Data Engineer. RTM SMEs and a compiler or interpreter specialist should provide domain and artifact guidance in either model.

Expected outputs

• A repeatable AWS document-ingestion and retrieval prototype with a known-answer evaluation scorecard.
• An inventory of available RTM compiler and interpreter artifacts, with evidence coverage and gaps documented.
• A versioned sample of structured RTM facts with source locations, extraction details, confidence, and SME review status.
• Deterministic query or API examples validated against selected SME-confirmed relationships.
• A concise technical handover covering configuration, run instructions, limitations, cost observations, and recommended next steps.