Yerassyl Manakbayev

@ManYera

3D Computer Vision · Data Analyst · Python Developer · ML/AI Engineer

3D Computer Vision · Data Analyst · Python Developer · ML/AI Engineer Bachelor's Degree, Applied Computer Science and Artificial Intelligence · Rome, Italy

Open to

Hiring & referralsStartup collaborationCompany partnerships

Career

Experience: 1 years

Experience

Research Intern, thesis project

VisionLab, Sapienza University of Rome

Sapienza is a public research university in Rome, founded in 1303, and one of Europe's largest universities by student enrollment.

Sep 2025 — Jul 2026

Built a 3D bone-fragment registration pipeline for a thesis project, reaching about 1 mm alignment accuracy across four real datasets and testing its robustness against GICP.

• Designed and implemented a complete 3D point cloud registration pipeline, from preprocessing and landmark-based initialization through IRLS-weighted ICP refinement and evaluation, using Python, Open3D, NumPy, SciPy, and Trimesh. • Achieved approximately 1 mm alignment accuracy when reconstructing bone fragments across four real datasets, combining FPFH feature extraction, coarse RANSAC alignment, and ICP refinement with robust Lp-norm weighting and surface-intersection constraints.

Education

Bachelor's degree, Applied Computer Science and Artificial Intelligence

Sapienza University of Rome

2023 — 2026

Selected projects

Tuberculosis Classification from X-rays

Built a ResNet18 classifier to distinguish tuberculosis-positive and normal X-rays, with about 95% training accuracy and 92% validation accuracy. Added U-Net lung segmentation for region-of-interest selection and Grad-CAM for interpretability. Trained and validated on the Kaggle TB Chest X-ray Dataset and Montgomery County Chest X-ray Set.

CS:GO Esports Match Analysis

Analyzed professional match data, including economy, outcomes, and player statistics, to assess the effect of winning the pistol round and cluster team play styles using k-means. Built a tournament dominance network analysis using centrality and community detection, and mapped the geographic distribution of players and teams.

WASAtext

Built a Go backend with a REST API and service layer for personal and group messaging, including media, emoji reactions, message forwarding, and read receipts. Containerized the backend and frontend with Docker for reproducible local deployment.

Skills

PythonJavaGoJavaScriptROpen3DOpenCVTrimeshPoint cloud registrationICPRANSACFPFHIterative alignmentSVDVoxel downsamplingPoisson disk samplingPolygon mesh processingNumPySciPyPyTorch