C++
Used for more than two years in a professional software development environment at IQVIA.
Grouped by evidence level: professional experience, own projects, current build focus, and semantic web background.
Tools and practices backed by the working student role at IQVIA.
Used for more than two years in a professional software development environment at IQVIA.
Used for debugging and working with an existing C++ application landscape.
Used for technical documentation, issue tracking, and work in established software workflows.
Used in professional collaboration and source control workflows.
Used as part of established development and CI processes in a professional software environment.
Applied in a working student role through functional analysis and clarification of software requirements.
Used to validate changes in an existing software system before handover or further review.
Used for documenting analysis, implementation context, and support-relevant information.
Technologies used in portfolio, Hainstreet, and other practical software projects.
Main language for backend and AI-related project work around FastAPI, RAG, evaluation, and automation.
Used and planned for Python backend APIs, health endpoints, typed request/response models, and service layers.
Used for safer frontend and backend-oriented web development in the portfolio and Hainstreet.
Used as a practical foundation for React, Next.js, Node.js, frontend behavior, and web application logic.
Used for reusable UI components, interactive product flows, forms, and portfolio pages.
Used for this portfolio and Hainstreet with app routing, server/client components, metadata, and deployment-ready pages.
Used as the runtime foundation for TypeScript services, development scripts, and backend-oriented web workflows.
Used in Hainstreet for dedicated backend services with modules, controllers, validation, and business logic.
Used in Hainstreet and planned as the main relational database foundation for the LLM reliability platform.
Used in Hainstreet as the ORM layer for schema management, generated types, and service-level data access.
Used for schema validation and safer request handling in TypeScript workflows.
Used for local development environments and deployment-oriented packaging in web and backend projects.
Used and planned for CI workflows, automated checks, builds, and project delivery steps.
Used for responsive layouts, utility-based styling, and consistent UI sections.
Used for browser-level checks and end-to-end validation of important user flows.
Used for JavaScript and TypeScript test coverage around application behavior.
Technologies and concepts currently being built into the llm-reliability-platform.
Current focus area for document-based AI assistants that retrieve relevant context before generating answers.
Being built for document ingestion, text extraction, chunking preparation, and later retrieval workflows.
Being built as a RAG building block for splitting documents into deterministic and retrievable text sections.
Being built to represent document chunks as vectors for semantic retrieval.
Being built for finding semantically relevant document chunks in RAG workflows.
Planned PostgreSQL extension for storing embeddings and running similarity search in the RAG system.
Being designed as an architecture layer that separates application logic from individual LLM providers.
Planned gateway behavior for selecting providers based on configuration, availability, and cost constraints.
Planned reliability mechanism for switching to fallback providers when the primary provider fails or times out.
Planned for making LLM usage visible across requests, providers, budgets, and evaluation runs.
Planned for checking answer quality, citation behavior, regression risk, and reliability over time.
Planned for protecting APIs, limiting demo usage, and reducing uncontrolled LLM cost risk.
Planned for tracing request flows across retrieval, provider calls, gateway behavior, and service boundaries.
Planned metrics system for latency, errors, token usage, fallback counts, and evaluation signals.
Planned dashboard layer for platform metrics, evaluation trends, provider behavior, and operational evidence.
Planned reverse proxy for HTTPS entry, routing, and simple deployment operations.
Planned hosting target for the live deployment of the LLM reliability platform.
Used and planned for DNS, proxying, TLS, and deployment-related infrastructure.
Planned for selected stateful LLM workflows such as retrieve, validate, gateway call, citation verification, and answer generation.
Relevant background from the award-winning CCC project and earlier development work.
Used in the award-winning CCC project to represent structured project cost knowledge.
Used in CCC for structured knowledge representation in a semantic web workflow.
Used in CCC for querying structured knowledge graph data.
Used in CCC and relevant for selected LLM workflow experiments in the LLM reliability project.
Used in the CCC question-answering project as part of the semantic chatbot workflow.
Familiar through university projects and object-oriented programming practice.
Used for version control, branching, commits, and daily development workflows.
Used for repository management, source code presentation, issues, project planning, and delivery workflows.
Used in this portfolio for contact-related email delivery and confirmation flows.
Additional background from earlier mobile and web development experiments.
Additional background from earlier native iOS development work.
Additional background from earlier work in the Apple development environment.