Amir Kamel Rahimi

Legal AI Agentic Search Assistant

• Architected and delivered a production RAG AI Assistant using Azure Functions and Azure AI Search. Implemented hybrid BM25 keyword and vector search, followed by Azure semantic reranking, with file-level filtering, relevance thresholds and duplicate removal to provide high-quality context for grounded LLM responses

Legal Document Processing & Summarisation Pipeline

• Built and deployed a recurring production pipeline using Azure Document Intelligence and Azure Cosmos DB to process legal case files, extract text through OCR, synchronise document content and metadata, and generate summaries for Slack canvases

AI Search Cost Optimisation

• Cut document data-extraction costs in Azure AI Search by at least 50% by re-designing the data-extraction workflow while maintaining retrieval quality. Evaluating local open-weight LLMs (gpt-oss, Qwen via Ollama) as a lower-cost alternative to cloud inference for very large documents

AI Evaluation & Content Safety

• Implemented RAG evaluation metrics, including faithfulness and answer relevancy, and content safety evaluations including unfairness in Azure AI Foundry

Data Integration & Automation

• Designed and deployed scalable, reusable data-ingestion and ETL workflows using Azure Document Intelligence and Azure Cosmos DB to extract, enrich and synchronise document content and metadata across internal systems

LLM Document Classification

• Designed and deployed AI classification model to automatically categorise legal documents to enhance case-routing capability

LLM Knowledge Graph Analytics

• Modelled clients’ financial data in a Cosmos DB knowledge graph, linking employment and payslip events to strengthen case analysis and evidentiary review

Executive Partnership

• Engage continuously with the CEO, CTO and partners to translate business needs into robust, enterprise-ready architectures that align AI delivery with strategic priorities

Technical Leadership

• Lead and mentor junior developers, conduct architecture, design and code reviews, and oversee deployment readiness, solution quality and alignment with engineering standards

Project "Smart Hub"

• Data Integration: Led the mapping of Queensland clinical data into the OMOP common data model using SQL, supporting scalable analytics and AI use cases. Supported the implementation of the native Gen AI model with GPT-2, trained by the OMOP standard data
• AI Solution Development: Contributed to the development of a domain-specific generative AI solution using PyTorch and standardised healthcare data
• Data Governance: Worked with clinical, technical and governance stakeholders to define data requirements and support secure, compliant access to sensitive healthcare data
• Stakeholder Collaboration: Translated complex clinical and technical requirements into practical data and AI delivery outcomes

Project "QChat"

• Enterprise GenAI Delivery: Designed and delivered QChat, a secure enterprise GenAI assistant using Azure OpenAI, Node.js, TypeScript and Next.js
• AI Governance & Safety: Implemented guardrails, monitoring and usage telemetry to support responsible AI adoption and production reliability across government teams

Project "Insight Engine"

• Gathered business requirements to develop a Gen AI solution aimed at guiding future strategic planning in QLD
• Designed and deployed a scalable RAG solution using Azure Functions, Azure AI Search and Azure Cosmos DB, ensuring scalable and efficient performance in line with future-focused strategic foresight
• Created knowledge graphs in NetworkX and applied clustering techniques (community detection) to create future strategic scenarios for QLD, contributing to both high-level foresight and actionable strategic directions

Project "Infection Prediction"

• Machine Learning Delivery: Led the design, validation and optimisation of an XGBoost model for predicting clinical infection risk using Python, Scikit-learn and MLflow
• Model Engineering: Applied feature engineering, Bayesian hyperparameter optimisation and model evaluation to improve predictive performance and reliability
• Cloud Deployment: Deployed the production machine-learning model using AWS SageMaker

Project "QUIET" (Queensland Integrated Element Tracker)

• Enterprise Software Engineering: Developed a .NET application supporting data governance, data modelling, metadata management and decision support across Queensland Health systems
• Database Design: Designed and maintained Azure SQL database schemas for enterprise healthcare data-management requirements
• Technical Design & Documentation: Produced technical specifications, architecture documentation, release notes and user guidance throughout the software-development lifecycle
• Engineering Quality: Contributed to code reviews, version control and collaborative delivery using GitHub and established software-engineering practices

Project "Genie Application"

• Full-stack developer on the Genie application in an Agile development environment
• Worked in agile development methodology for writing code, test units and code maintenance

Project "Legacy Data Migration to AWS"

• Lead data migration engineer, responsible for migrating legacy clinical data to Amazon Cloud
• Mentored junior developers, providing technical guidance and support on development best practices, code reviews, and quality assurance

• Worked as a research assistant, focusing on the development and validation of a comparison application for DNA sequences in Java
• Conducted extensive research on various algorithms and techniques for assessing DNA sequences to identify patterns and variations
• Collaborated with the research team to design and implement the software tool, ensuring accuracy, scalability, and performance for analysing large datasets of DNA sequences