GitHub • LinkedIn • Hugging Face • Mail
I work on AI systems that combine data, vision and language.
My focus is on:
- building end-to-end machine learning pipelines
- working with real-world data
- designing search, retrieval and recommendation systems
- applying NLP & computer vision together
- experimenting with generative AI and LLM-based systems
I enjoy turning research ideas into working, testable, and scalable systems.
A production-ready AI system that understands products using both text and images, with GenAI features including RAG pipeline and AI agents.
What it includes:
- Semantic text search
- Image-based similarity search
- Text–image fusion (learned α=0.7)
- Vector search with FAISS
- RAG pipeline with LLMs
- AI agent with conversation memory
- Personalization system
- Ranking & evaluation (NDCG, Recall, MRR)
- Full-stack application (React + FastAPI + MongoDB)
- End-to-end ML pipeline
Results:
- 97.4% NDCG@10 (search performance)
- SUS 84.50 (Grade A) - user study with 25 participants
- 92% real-world usage intent
- 100% AI agent success rate
Tech used:
- CLIP, Sentence Transformers
- FAISS, LightGBM
- GROQ (Llama-3.3-70B), LangChain
- FastAPI, React, MongoDB
📌 Repository:
👉 https://github.com/haticebaydemir/ai-fashion-assistant-v2
- Supervised / Unsupervised Learning
- Representation Learning
- Multimodal Models
- Generative AI (RAG, AI Agents, LLMs)
- Model Evaluation
- Image embeddings
- Visual similarity search
- CLIP-based vision–language models
- Image processing
- Semantic search
- Sentence embeddings
- Text similarity
- Query understanding
- LLM integration & prompt engineering
- FAISS vector search
- Dense retrieval
- Ranking & evaluation
- Hybrid search systems
- Retrieval-Augmented Generation (RAG)
- FastAPI, React
- MongoDB, PostgreSQL
- JWT authentication
- Deployment (Hugging Face, Docker)
- Multimodal AI
- Semantic search systems
- Generative AI (RAG, Agents)
- Retrieval-augmented pipelines
- Applied machine learning
- Full-stack AI applications
Building intelligent systems with data, vision and language.