Skip to content
View Mervecaliskann's full-sized avatar
14 followers · 46 following

Block or report Mervecaliskann

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Mervecaliskann/README.md

Hi there, I'm Merve! 👋

🚀 Computer Vision & AI Engineer Master's Student in Data Science | Specializing in End-to-End AI Pipelines, Computer Vision, and Production-Ready LLM Applications.

  • 💡 Background in Electrical & Electronics Engineering.
  • 🎓 Pursuing MSc in Data Science @ Dokuz Eylül University (GPA: 3.70)
  • 🏆 1st Place — LLM Jetpack Ideathon | AI & CV Engineer @ Glossa (Pre-seed Beauty-Tech)
  • 🌍 Top 10 Talent — EU-funded E-DATA Project (Netherlands & Germany)

🛠️ Tech Stack & Tools

🔬 Featured Projects

Project Description Stack
🦙 Llama-3.2-3B QLoRA Fine-Tuning Fine-tuned Llama-3.2-3B on Dolly-15K. Train loss 1.87→1.49, no overfitting. PyTorch, Unsloth, TRL, MLflow
🤖 Transformer from Scratch Full encoder-decoder: MHA, Beam Search, FP16, LR Scheduler, SST-2 PyTorch
📊 AI Agentic Benchmarking Lab ReAct vs Plan-and-Execute benchmarking with statistical validation BigQuery, SciPy, Streamlit
🎫 AI Support Analytics Pipeline End-to-end ticketing: FastAPI + Llama-3 + PostgreSQL + CI/CD FastAPI, Groq, Docker
🎨 Glossa Color Correction Engine CIELAB-based metamerism fix + Digital Grey Card for beauty-tech OpenCV, CIELAB, Flutter

🌱 Current Focus

  • Advanced RAG Architectures (Agentic & Multimodal workflows).
  • MLOps: Scaling and deploying LLMs & CV models in production environments.

📫 Let's Connect:

Pinned Loading

  1. AI-Support-Pipeline AI-Support-Pipeline Public

    An end-to-end AI support pipeline that automates ticket categorization, sentiment analysis, and priority assignment using FastAPI, Llama-3, and PostgreSQL with a real-time Streamlit dashboard.

    Python

  2. AI-Data-Analyst AI-Data-Analyst Public

    A hybrid AI agent for automated data cleaning. Combines Pandas for deterministic statistical profiling and Llama-3 for probabilistic reasoning and Python code generation.

    Python

  3. AI-Finance-Agent AI-Finance-Agent Public

    Real-time AI Financial Analyst Agent capable of fetching live stock market data and performing fundamental analysis using Tool Calling. Built with LangChain, Groq (Llama-3.3), and Yahoo Finance API.

    Python 1

  4. Multi-Source-RAG-Application-with-LangChain-Google-Gemini Multi-Source-RAG-Application-with-LangChain-Google-Gemini Public

    This project demonstrates a powerful Retrieval-Augmented Generation (RAG) application built using Python, LangChain, and Streamlit. It enhances the capabilities of Google's Gemini LLM by providing …

    Python 1

  5. agent-benchmark-pipeline agent-benchmark-pipeline Public

    A statistical R&D framework for benchmarking LLM Agent architectures (ReAct vs. Plan-and-Execute). Features automated Google BigQuery logging and an adaptive statistical layer (ANOVA/Kruskal-Wallis…

    Python

  6. flight-delay-analysis-R flight-delay-analysis-R Public

    Havalimanı Büyüklüğü, Mevsimsel Faktörler ve Uçuş Rotası Özelliklerinin Uçuş Gecikmeleri Üzerindeki Etkisinin İncelenmesi ve Bu Faktörler Arasındaki İlişkilerin Analizi