Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I'm a recent graduate of Duke University with a Bachelor of Science in Computer Science. My research focused on artificial intelligence, specifically in computer vision models, where I gained extensive experience with classification, segmentation, and feature matching. Since graduating, I've been applying this practical knowledge at an AI startup in Tbilisi, where I'm developing a mobile human pose estimation application. In this role, I have specialized in translating visual input into 3D spatial data, a direct parallel to the project's requirement to extract 3D information from 2D imagery. My experience includes end-to-end product development, from initial research and model training to deployment and stakeholder communication, which I believe makes me a strong fit for your project.…I'm a recent graduate of Duke University with a Bachelor of Science in Computer Science. My research focused on artificial intelligence, specifically in computer vision models, where I gained extensive experience with classification, segmentation, and feature matching. Since graduating, I've been applying this practical knowledge at an AI startup in Tbilisi, where I'm developing a mobile human pose estimation application. In this role, I have specialized in translating visual input into 3D spatial data, a direct parallel to the project's requirement to extract 3D information from 2D imagery. My experience includes end-to-end product development, from initial research and model training to deployment and stakeholder communication, which I believe makes me a strong fit for your project.

Nikoloz Kiladze

AI Engineer, App Developer, Programmer, +1





I'm a recent graduate of Duke University with a Bachelor of Science in Computer Science. My research focused on artificial intelligence, specifically in computer vision models, where I gained extensive experience with classification, segmentation, and feature matching. Since graduating, I've been applying this practical knowledge at an AI startup in Tbilisi, where I'm developing a mobile human pose estimation application. In this role, I have specialized in translating visual input into 3D spatial data, a direct parallel to the project's requirement to extract 3D information from 2D imagery. My experience includes end-to-end product development, from initial research and model training to deployment and stakeholder communication, which I believe makes me a strong fit for your project.…I'm a recent graduate of Duke University with a Bachelor of Science in Computer Science. My research focused on artificial intelligence, specifically in computer vision models, where I gained extensive experience with classification, segmentation, and feature matching. Since graduating, I've been applying this practical knowledge at an AI startup in Tbilisi, where I'm developing a mobile human pose estimation application. In this role, I have specialized in translating visual input into 3D spatial data, a direct parallel to the project's requirement to extract 3D information from 2D imagery. My experience includes end-to-end product development, from initial research and model training to deployment and stakeholder communication, which I believe makes me a strong fit for your project.

Available to hire

I’m a recent graduate of Duke University with a Bachelor of Science in Computer Science. My research focused on artificial intelligence, specifically in computer vision models, where I gained extensive experience with classification, segmentation, and feature matching. Since graduating, I’ve been applying this practical knowledge at an AI startup in Tbilisi, where I’m developing a mobile human pose estimation application. In this role, I have specialized in translating visual input into 3D spatial data, a direct parallel to the project’s requirement to extract 3D information from 2D imagery. My experience includes end-to-end product development, from initial research and model training to deployment and stakeholder communication, which I believe makes me a strong fit for your project.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Docker

Intermediate

Language

Georgian

Fluent

English

Fluent

Russian

Fluent

Italian

Intermediate

Work Experience

Mid-level AI Engineer at Arrrteo – AI Startup

August 31, 2025 - August 26, 2025

Engineered a lightweight, mobile-optimized computer vision model for human pose estimation that translates real-time mobile camera input into 3D joint positions at 20 FPS within a Unity environment. Conducted research on state-of-the-art models for head and full-body tracking, managing the entire training and testing pipeline to ensure high accuracy with a mean per-joint position error of 54.56 mm. Developed the final model in Python and converted it to ONNX for integration in Unity. Presented AI project progress and strategic roadmaps to stakeholders, effectively bridging technical and business teams.

Mid-level AI Engineer at Arrarteo – AI Startup

August 31, 2025 - August 26, 2025

Engineered a lightweight mobile-optimized computer vision model for human pose estimation, translating real-time mobile camera input into 3D joint positions at 20 FPS for Unity integration. Led research on state-of-the-art models for head and full-body tracking, managing end-to-end training and testing pipelines with an Overall Average Error (MPJPE) of 54.56 mm. Developed and refined the model in Python, converting it to ONNX for seamless Unity implementation. Presented AI project progress and strategic roadmaps to key stakeholders, effectively bridging technical and business teams.

Student Research Assistant at Duke University (Dr. Xiaowei Yang, Professor of Computer Science)

May 31, 2025 - August 26, 2025

Trained three distinct CNNs for image classification, segmentation, and node connection analysis on network topology maps using Slurm computing resources. Developed a web crawler and classifier to process approximately 2,000 images from 20 websites with 90% accuracy. Created and curated a large-scale dataset of 1000 images through automated pipelines, conducted model evaluation and benchmarking for performance consistency. Transformed processed images into graph formats and applied LoFTR for real-world location matching.

Student Research Assistant at Duke University - Uncovering Cloud Network Topologies

May 31, 2025 - August 26, 2025

Trained three distinct CNNs using Python for image classification, segmentation, and node connection analysis on network maps with Slurm computing resources. Developed a web crawler and trained classifier to process ~2,000 images from 20 websites with 90% accuracy. Curated a large-scale image dataset of 1,000 images with automated pipelines; conducted evaluation and benchmarking across multiple CNN architectures. Transformed imagery into graph formats and applied LoFTR for real-world location matching.

Senior Thesis Developer at MapForge: AI-Enhanced Geospatial Data Visualization Platform

April 30, 2025 - August 26, 2025

Developed an open-source, user-friendly web-based platform using React.js, Mapbox GL JS, and Express.js to streamline spatial data acquisition and visualization for researchers. Integrated NLP with the Anthropic Claude API, enabling natural language queries with over 85% accuracy. Engineered backend prompt pipelines reducing hallucinations by 25% for improved factual consistency. Built dynamic data tools including feature grouping, customizable map styling, and support for GeoJSON and CSV uploads, tested on 10+ real-world spatial datasets from environmental and urban studies.

Co-founder at Cura: AI-Powered Cancer Literacy and Risk Prediction Startup

August 31, 2024 - August 26, 2025

Co-developed an iOS application focused on increasing cancer literacy using Swift. Designed and evaluated an NLP-driven chatbot feature leveraging seven different large language models with fine-tuning to provide personalized cancer-related guidance. Collected and processed 183 patient records to train a logistic regression cancer risk prediction model achieving 88.5% accuracy.

Co-founder at Cura: AI-Powered Cancer Literacy and Risk Prediction Startup

August 31, 2024 - August 26, 2025

Co-developed an iOS application focused on increasing cancer literacy using Swift. Designed and evaluated an NLP-driven chatbot employing seven different LLMs with fine-tuning for personalized cancer-related guidance. Collected and processed a dataset of 183 patient records (25 attributes each) to train a logistic regression cancer risk prediction model achieving 88.5% accuracy.

Student Research Assistant at Duke University (Dr. Ashutosh Kotwal, Fritz London Distinguished Professor of Physics)

May 31, 2024 - August 26, 2025

Collaborated within a team of 10 to reconstruct charged particle trajectories for LHC experiments. Translated approximately 2,500 lines of C++ code to C for hardware synthesis using VITIS.

Student Research Assistant at Duke University - Searching for Metastable Particles Using Graph Computing

May 31, 2024 - August 26, 2025

Collaborated in a team of 10 reconstructing charged particle trajectories for LHC experiments. Translated approximately 2,500 lines of C++ code to C for hardware synthesis using VITIS.