Mert Ege Email & Phone Number

Ankara, Türkiye

- I am the technical lead of the pretraining and supervised fine-tuning processes of the LLMs.- I am on the LLM team as an LLM/Research Engineer.- As a team, we are developing a multi-turn question-answering model for customers using GPT-based LLMs.- For fine-tuning tasks, I am leading five researchers. By the way, I am in contact with other departments (evaluation, db, mini-models) to construct the end-to-end system.- I am using Python/PyTorch to build models.- Our team includes Turkish, Arabian, and Chinese researchers. So, we communicate in English.Skills: PyTorch · NLP · GPT · LLM · Pretraining · SFT · Retrieval Augmented Generation (RAG) · Python

İstanbul, Türkiye

- I work in the Query Correction module of the Petal Search project for the Arabic Language.- I'm researching and implementing machine learning/deep learning algorithms for the NLP project Petal Search at Huawei.- Machine learning models, like CRF, Catboost, Ngram-based, and SVM, are used for this task. Furthermore, I am working on implementing Large Language Models (LLMs) like BERT, T5, and GPT-based models.- I created and implemented a sequence-to-sequence Large Language model (T5) for the query correction module.- I have multicultural (Turkish, Indian, Arabian, and Chinese) teammates in this project.

İstanbul, Türkiye

I was responsible for the AI modules in the Talent Interview project.Online interviews are held in the Talent Interview project.My task in this project was to automatically detect cheating.So, we used different methods to detect cheating like Mobilenet Coco-SSD, Body-Pix, and Speech Commands.Technologies: Python, JavaScript, Tensorflow, Tensorflow JS, Keras, Mobilenet Coco-SSD, Body-Pix, and Speech Commands.Domains: Object Detection, Human Voice Detection, Computer Vision, Image Segmentation, and Web-Side ML.

Ankara, Türkiye

I worked as a grader in different courses.

Ankara, Turkey

I was working as a digital design engineer at ASELSAN. I was doing Computer Vision operations on thermal and visible image data using FPGA chips that can perform parallel processing with low power consumption. I was performing these operations using the VHDL language.

Ankara, Turkey

Ankara, Türkiye

Aydın, Türkiye

Master Of Science - Ms, Electrical And Electronics Engineering

Thesis Title: Human Activity Classification with Deep Learning using FMCW Radar

Bachelor Of Science - Bs, Elektrik Ve Elektronik Mühendisliği

Graduated with Honor Degree

Education record

Education record