Kevin Bals

AI Engineer / Backend GenAI Engineer at Accenture Industry X, I work on industrial innovation projects involving generative AI, RAG, LLM agents, and document intelligence.

My role involves designing, developing, and industrializing production-oriented GenAI solutions: Python/FastAPI backends, document ingestion pipelines, PDF extraction and structuring, hybrid search, LLM orchestration, AI agents, sourced answers, and cloud deployment on GCP.

I have contributed to several AI platforms for demanding industrial environments, particularly documentary assistants capable of querying complex technical corpora with reliable, traceable, and sourced answers.

Key Responsibilities:

- Python backend development with FastAPI, Pydantic, SQLAlchemy, and REST APIs;

- Design of RAG systems with hybrid search, embeddings, PostgreSQL, pgvector, BM25, and pg_trgm;

- LLM agent orchestration with LangGraph, Google ADK, Gemini, and Vertex AI;

- Document extraction with PyMuPDF, Camelot, OCR, and Google Document AI;

- Deployment and industrialization on GCP: Cloud Run, Cloud SQL, Firestore, GCS, Pub/Sub, Docker, Terraform, and CI/CD;

- Collaboration with business and technical teams to transform complex industrial needs into usable AI solutions.

This experience has allowed me to develop solid expertise in building reliable, maintainable, and integrable GenAI systems in real cloud environments.