GDG Meetup – Building Knowledge Graphs from Unstructured Data (Ciklum, 2024)
Info
- Project type Tech meetup, Live coding
- Participants 35
- Date 24th of Sept 2024
- Duration 1h
- My Role Speaker
- Topics Knowledge Graphs, Neo4j, Vertex AI, Gemini
- Keywords LLMs, Graph DB, Knowledge representation, Graph RAG, Unstructured
- Skills developed Public speaking
- Presentation Slides
-
Photo
Description
In Sept 2024, I delivered my second talk at Google Developer Group Bucharest titled "Building Knowledge Graphs from Unstructured Data", an event organized by Digital Stack and Ciklum Romania.
The talk introduced fundamental concepts about Graph RAG and demonstrated how this paradigm overcomes the limitations of vanilla RAG, using real-life examples and a live coding demo.
My audience consisted of students and developers passionate about AI and software architecture.
The talk's brief:
RAG (Retrieval-Augmented Generation) has been one of the most talked-about topics in AI this year, with many organizations pioneering its implementation. While significant advancements have been made in data chunking strategies and retrieval techniques, the traditional RAG architecture still faces limitations due to its reliance on vector stores for knowledge bases.
In this workshop, we'll explore how integrating Knowledge Graphs into RAG can overcome these limitations and elevate the power of understanding in AI applications.
In this workshop, we will explore:
- Processing highly complex and unstructured data using Gemini API and Vertex AI
- Understanding the significance of entities and relationships in Graph Theory
- Practical tools for generating and working with Knowledge Graphs
- The advantages of using Graph Databases in RAG over traditional relational and vector databases
This session is specifically addressed to developers and data enthusiasts who are eager to explore the next level of innovation with RAG and knowledge-based AI systems.
During the workshop, I demonstrated live coding and use cases I encountered while building my framework, ContinualBot.com.
I thoroughly enjoyed my time at the meetup, especially the engaging questions during the Q&A session and the quality networking after the event.
I also appreciated the commitment of the GDG to organizing such professional events, this time with the support of Ciklum. This was an opportunity to expand my community and network connections in the Romanian software development community, where I had the pleasure of meeting many intelligent and inspiring individuals.
This meetup was particularly memorable for me because my co-speaker was my long-term collaborator and friend, Tere. We have worked together on many software projects since our student days at ETTI, starting with leadership roles in LSE's IT Department and continuing with projects like MobilPRO, CareerTeam, TEDxUPB, Robochallenge, TuneGang, Ionadi, Trade & Created in Romania, Questo, Cuibul Artiștilor. It is amazing to cross our ways again and explore cutting-edge technologies together.
Tech stack
- Programming Languages: Java, Python, JavaScript, SQL, Cypher
- Databases: MariaDB, Qdrant, Neo4j
- LLMs: Google Gemini, OpenAI GPT
- OS: Windows, Linux
- Cloud: GCP, Vertex AI, Google AI Studio