AI Engineer at Independent (Remote)

Develop private GenAI and local RAG solutions for freelance projects, prioritizing data privacy, infrastructure control and lower inference costs through local LLM deployments. Built a local AI assistant for documentation workflows using an OpenAI-compatible API, PostgreSQL/pgvector, streaming responses and conversational memory. Designed agent-based workflows for retrieval, routing and answer generation, focused on practical B2B use cases and production-ready integration.

AI Engineer at Wembley Studios

Built a B2B purchasing assistant by unifying ~20 PDF catalogs with SQL data (~19k products and ~5k spare parts), creating enriched datasets for conversational retrieval. Engineered a RAG pipeline on Azure AI Search with hybrid retrieval, semantic ranking and multi-query retrieval to improve recall on ambiguous product queries. Architected a tool/function-calling router to direct queries across products, spare parts, catalogs and history sources, improving answer accuracy and reducing unnecessary calls. Automated the processing of ~500 invoices per month with Azure Functions and Document Intelligence, including edge-case retraining and fuzzy matching against database records for installer validation. Developed semantic matching between invoice lines and indexed catalog content, combining embeddings with price and keyword heuristics to improve reliability on noisy invoiced data. Built a Microsoft Teams support bot that retrieved indexed incidents and created tickets via API when no suita

Freelance AI Engineer (Remote) at Independent / Freelance

Develop private Gen AI and RAG solutions for freelance projects; prioritized data privacy and infrastructure control. Built a local AI assistant for end-user documentation and workflows using an OpenAI-compatible API, PostgreSQL (pgVector), streaming responses and conversational memory. Designed agent-based workflows for retrieval, routing and answer generation, focusing on practical B2B use cases and production-ready integration.

AI Engineer at Wembley Studios (B2B)

Built a B2B purchasing assistant by unifying ~20 PDF catalogs (10-200 pages) with SQL data (pgvector), creating enriched datasets for conversational retrieval. Engineered a RAG pipeline on Azure AI Search with semantic ranking and multi-query retrieval to improve recall on ambiguous product queries. Architected a tool/function-calling router to direct queries across products, spare parts, catalogs and history sources, improving answer accuracy and reducing unnecessary calls. Automated processing of ~500 invoices per month with Azure Functions and Document Intelligence, including edge-case retraining and fuzzy matching against database records for installer validation.

Develop private Gen AI and local RAG solutions for freelance projects, prioritizing data privacy, infrastructure control and lower inference costs through local LLM deployments; built a local AI assistant for end-to-end documentation and workflow using an OpenAI-compatible API, PostgreSQL/pgvector, streaming responses and conversational memory; designed agent-based workflows for retrieval, routing and answer generation, focused on practical B2B use cases and production-ready integration.