Looking for a job

Giovanni wang

@giovanniwang

Message

I build production RAG systems, search platforms, and scalable Java/Python backends.

Italy

Message

What I'm looking for

I'm looking for fully remote work where I can own production AI, RAG, search, or backend services end to end in a distributed, async-first team with strong written handoffs, measurable quality, and minimal oversight.

At Lexroom, I built traceable RAG and agent systems over an 80,000-document Italian legal corpus, connecting generated answers to supporting source passages.

I replaced vector-only retrieval with hybrid PostgreSQL/pgvector and BM25 search, Reciprocal Rank Fusion, and cross-encoder reranking, raising recall@5 from 0.68 to 0.87. I also built CI-gated evaluation harnesses and LangGraph agents that improved faithfulness from 0.85 to 0.93.

Previously, I delivered Python/FastAPI data and search platforms at Lexroom and Fairmat, including resilient ingestion pipelines, enterprise connectors, ACL-aware retrieval, PostgreSQL FTS, Elasticsearch, Celery, Docker, and AWS deployments.

At Restworld, I owned Java/Spring Boot marketplace services and improved candidate matching, search latency, integrations, and caching. I’m comfortable owning services end to end in distributed, async-first teams.

Experience

Work history, roles, and key accomplishments

AI Engineer — RAG, LLM & Agents

Lexroom S.r.l.

Jan 2025 - May 2026 (1 year 4 months)

Designed the RAG chain end to end, including chunking, embeddings, pgvector retrieval, and inline citations. Built a CI-gated evaluation harness and shipped multi-step LangGraph agents with tool calling, improving faithfulness and reducing unsupported answers.