Hey! I am

Gurami Keretchashvili

I'm a

About

About Me

I’m an AI Engineer with over six years of experience working in computer vision, natural language processing, and signal processing. I have a Master’s degree in Artificial Intelligence from Institut Polytechnique de Paris, one of the top universities in the world. As a certified PMI-Agile Practitioner (PMI-ACP), I focus on working smarter and leading teams effectively to deliver high-quality results. I enjoy using AI to solve real-world problems and always look for ways to learn more about new technologies. I also like sharing what I know and helping others grow in the field.

  • Name: Gurami Keretchashvili
  • Date of birth: April 11, 1998
  • Address: Paris, France
  • Email: g.keretchashvili@gmail.com
  • Phone: +33(0)7 66 11 25 43

0 Project complete

Download My Resume

Education

2020 - 2022

Master of Science in Artificial Intelligence

Institut Polytechnique de Paris

- GPA: 14.5/20 (Grade: A)

- Master Thesis: Borehole Sonic Data Classification.

- One of the top 50 universities in the world.

2016-2020

Bachelor of Science in Computer Engineering

San Diego State University

- GPA: 3.48/4
- member of dean's list several times

2010-2016

High School Diploma

Tbilisi I. Vekua Physics-Mathematical School 42

- GPA: 9.9/10
- a member of the group preparing for the National Olympiad in Physics and Mathematics
- a group leader, responsible for extra curriculum activities.

Experience

Jan 2023 - Present

AI Engineer

SLB

- Leading a deep learning science team
- Designing and building deep learning models for computer vision, NLP, and time series tasks
- Working with stakeholders to plan projects, understand needs, and lead the work
- Mentoring interns and helping them grow their skills
- Giving talks and presentations at webinars
- Exploring new AI tools and ideas to improve our work
- Writing research papers and working on patents.
- Created predictive models and data visualizations.
- Developed and advised data-driven solutions.

Mar 2022 - Sept 2022

AI Researcher intern

Veon

- Master Thesis: Borehole Sonic Data Classification using ML

Sep 2019 - Jan 2022

Data Scientist

Veon

• Created predictive models and data visualizations
• Consulted on effective data-driven solutions

Jan 2019 - Feb 2020

Teaching Assistant in Embedded System Programming

San Diego State University

• Guided undergraduate students through laboratory experiments
• Programmed AVR/ARM microcontrollers using C language

Jul 2018 - Aug 2018

Team Leader

Youth Agency of Georgia

• Led up to 25 students aged 15 to 22 in a summer camp
• Organized summer camp activities

Oct 2016 - Jun 2018

Mathematics & Physics Teacher

Auditoria #1

- Prepared up to 80 high school students in Mathematics and Physics for their final exam.
- All my students passed their exams successfully

Feb 2017 - Mar 2017

Physics Problem Writer

murtsku.com

- Helped high school students to prepare for national exams (NAEC) in math and physics

Skills

- Machine Learning and Deep Learning Algorithms

- Python (Pandas, Matplotlib, Sklearn, NetworkX)

- Data Visualization

- Natural Language Processing (NLP)

- Tensorflow (Google Certified)

- SQL, Database Management Systems

- Graph Mining

- Image Processing

- Project Management - PMI Agile Certified Practitioner (PMI-ACP)

- IT Project Management

Awards

Feb 2023

Winner of Data Challenge (Georgian Kaggle) - NLP (Team: Synapse NLP)

Bank of Georgia

• 3nd place out of more than 100 teams.
• Created sentiment analysis classifier using SOTA deep learning algorithms.
Article

May 2022

Winner of Data Challenge (Georgian Kaggle) - Recommender System (Team: Synapse)

Bank of Georgia

• 2nd place out of 120 teams
• Created recommender system algorithm using SOTA deep learning algorithms.
Article

Dec 2019

Winner of Datathon Competition (TEAM: Unofficial Intelligence)

TBC bank

• Winner in insurance business category.
• Created a churn prediction model.
Article

Datathod
Mar 2019

Winner of Big Data Hackathon in San Diego (TEAM: FAME)

Big Data Hackathon(San Diego)

• Winner of Aging Independently category.
• Predicting their activities based on specific time.
Article

big data hackathon
May 2016

Winner of Logic Game competition

caucasus university

• Competition in Mathematical Logical problems •
2000Lari scholarship
Article

Logic Game competition
Apr 2016

Finalist of Annual Georgian Olympiad in Mathematics and physics

georgia's national olympiad

• One of the 50 finalist in Georgia

Blog

My Blogs

Sharing knowledge and experiences in AI, Technology, and Personal Development
Follow me on Medium

AI & Machine Learning

Personal Development & Productivity

Publications

My Publications

Research papers and academic contributions in AI and Machine Learning

Revolutionizing Well Integrity: Temporal Deep Learning for Precise and Continuous Axial Groove Detection

May 17, 2025 SPWLA

This paper proposes TAUNet, a spatio-temporal model using bidirectional convolutional LSTMs to detect corrosion defects in wellbore images by capturing spatial and temporal continuity. It introduces temporal entropy loss and temporal bridging postprocessing to ensure consistent, continuous predictions. Custom data augmentation further improves generalization. The framework significantly boosts segmentation accuracy and could benefit other fields needing temporal coherence.

Machine Learning-Enabled Joint Interpretation of Dipole Sonic and Borehole Image Data

Jun 10, 2023 SPWLA

The paper introduce machine learning workflow which can automatically identify many important geomechanical and geological features, such as breakouts, natural fractures, drilling fractures, stress effects and intrinsic anisotropic effects using borehole sonic and image data with the help of machine learning.

Combining Embeddings and Rules for Fact Prediction

May 25, 2022 AIB

This paper proposes a survey of neuro-symbolic approaches that combine Rule Mining and Knowledge Graph Embeddings to predict missing facts in incomplete knowledge bases. Rule Mining discovers logical patterns, while Knowledge Graph Embeddings map entities to vector spaces. By integrating these methods, the paper aims to leverage their complementary strengths for more accurate fact prediction.

I'm Available for freelancing

Contact Me

Contact

Contact Me