Ifon

Key Skills

Software

No software listed

Top Subject Matter

AI Training $ Evaluation

Prompt Engineering

RAG & Knowledge Graph Systems

Top Data Types

Document

Computer Code Programming

Image

Top Task Types

Text Generation

Relationship

Question Answering

Computer Programming/Coding

Segmentation

Classification

Entity (NER) Classification

Text Summarization

Data Collection

Evaluation/Rating

Function Calling

RLHF

Object Detection

Bounding Box

Polygon

Point/Key Point

Polyline

Fine-tuning

Prompt + Response Writing (SFT)

Transcription

Red Teaming

Freelancer Overview

Master’s graduate in Computer Science from University of Amsterdam and Vrije Universiteit Amsterdam with research and hands-on experience in AI training data, LLM evaluation, and data quality workflows. Brings 1+ years of professional experience across complex professional workflows, research, quality-focused execution, and retrieval-augmented generation (RAG). My technical strengths include LLM prompting, relation extraction, vector retrieval, and annotation-oriented workflow design using tools such as Ollama, FAISS, Neo4j, Sentence Transformers, and Docker. Through both academic research and applied projects, I have developed expertise in text generation, semantic labeling, document understanding, and relationship annotation. I am particularly interested in AI training workflows involving complex reasoning, structured knowledge extraction, and quality evaluation for LLM systems.

Labeling Experience

Collaborative Project with Tongji University - RAG Question-Answering System for Periodontal Disease Knowledge Graph

DocumentQuestion Answering

Developed a RAG-based question-answering system for periodontal disease knowledge using an integrated Neo4j knowledge graph. Built a pipeline that extracts entity-relation-entity (ERE) triplets from unstructured PDF medical literature and supports relation-aware retrieval for downstream LLM question answering. Implemented semantic retrieval using Sentence Transformers and FAISS to improve relevance and response quality.

2025 - 2025

Master Thesis - Efficient Table Error Detection with LLMs

TextData Collection

Conducted research on LLM-based table error detection and data quality assessment as part of a Master’s thesis on efficient table error detection with large language models. Iterated on zero-shot and few-shot prompting strategies to improve F1 performance over baseline methods. Implemented clustering-based sampling and batch prompting to reduce token usage while maintaining detection quality.

2025 - 2025

Education

Work History