Open to opportunities

Ivan Klimovich

@ivanklimovich

Message

Junior Python developer building neural networks for biological image analysis.

Belarus

Message

What I'm looking for

I’m looking for a junior role where I can develop Python and machine learning for computer vision in biology—building reliable pipelines, publishing useful tools, and collaborating in OpenSource with real-world impact.

I am a 3rd year student of the Biology Faculty of BSU with advanced self-study in programming and machine learning. I develop neural networks for analyzing images of biological objects in Python, and I’ve published scientific articles on ML and Python, including a package for adaptive image binarization on PuPi.

I’m an active participant in OpenSource, and I build practical tools for biomedical imaging. In BioThreshold, I created an adaptive segmentation library for complex biological images using methods like Niblack, Sauvola, and local Otsu, integrating deep learning via Cellpose; it reduced segmentation errors by 25%. In my current work, I’m training a CNN for phenomic analysis of Chlorella vulgaris with PyTorch and ResNet-18, achieving IoU = 0.84 ± 0.05 and Dice = 0.89 ± 0.04.

Experience

Work history, roles, and key accomplishments

Current

Chlorella Vulgaris Cell CNN

Current

Chlorella Vulgaris CNN

Sep 2025 - Present (9 months)

Built and trained a PyTorch CNN for automatic segmentation and quantitative analysis of Chlorella vulgaris microalgae from micrographs. Used ResNet-18 transfer learning with a two-head architecture for simultaneous regression of morphology (area) and color (RGB), achieving IoU 0.84 ± 0.05 and Dice 0.89 ± 0.04.

ResNet18 PyTorch Computer Vision OpenCV Scikit Image Transfer Learning Regularization (Dropout BatchNorm)

BioThreshold Adaptive Segmentation

BioThreshold

Jan 2025 - Apr 2026 (1 year 3 months)

Developed BioThreshold, a Python library for adaptive threshold segmentation of complex biomedical images (algae, plant calli, histology). Reduced segmentation errors by 25% using CLAHE + Niblack, added configurable preprocessing/postprocessing, and integrated Cellpose for deep-learning-assisted segmentation.

Python OpenCV Scikit Image NumPy Segmentation CLAHE Thresholding (Niblack Sauvola)Cellpose Integration

Education

Degrees, certifications, and relevant coursework

Belarusian State University

Bachelor's degree (in progress), Biology

Activities and societies: Active in open source; published scientific articles on ML and Python; develops ML tools for biological image analysis.

3rd-year student at the Biology Faculty of Belarusian State University, combining biology studies with advanced self-study in programming and machine learning. Builds neural networks for analyzing images of biological objects and has published ML/Python work.