Hi! My name is Viktor and I am currently a software engineer at Caterpillar. In my free time I work on a couple of side projects, bake bread, and do ballroom dancing.
Viktor Chekhovoi
Student, Carleton College
Hi! My name is Viktor and I am currently a software engineer at Caterpillar. In my free time I work on a couple of side projects, bake bread, and do ballroom dancing.
As a full-time software engineer, I developed and implemented the AWS architecture to optimize a core feature of Caterpillar's Equipment Condition Monitoring. For this, I used Apache Spark and EMR Serverless to retrieve and analyze usage statistics for 1.6 million assets, and calculate the likelihood of an asset needing repairs. My solution cut down run time from 6 hours to 30 minutes, and reduced monthly costs for processing the data to $30. I implemented an autoscaling solution for my team's databases to accomodate spikes in read/write requests, helping prevent several outages. I also created an internal database versioning tool using Flyway that's now used on several internal databases.
I contributed to the development of Foresight, a risk prediction service for Caterpillar clients that collects data from Caterpillar assets and uses multiple machine learning models to predict when a certain piece of equipment is going to break or need repairs. I developed a key feature for an early Foresight release, a filter engine for removing false positive risk predictions. For that project, I had a leading role in the development of an AWS ECS service with an Autoscaler, and created a Dockerized application for it. In another project, I led the development of a workflow for processing machine learning model output using AWS ECS, Apache Spark and Hudi. My solution used an innovative storage format to utilize AWS S3 instead of DynamoDB, leading to an up to 70% reduction in storage costs for infrequently accessed data. I also regularly worked with CI/CD pipelines and datalake software such as Snowflake.
As an Academic Technology Assistant, my responsibilities include providing support for Moodle and Gather Town, as well as monitoring and addressing issues with Carleton's software. I also contributed to the database solution for Carleton's Language Lessons - a web application for learning languages.
Carleton College is ranked #21 among liberal arts colleges in the US. I graduated with a double major in Computer Science and Math, and received a distintion in Computer Science for my final project.
Kyiv Natural Science Lyceum is one of the best schools in Ukraine, ranked #1 in 2015. Ever since it was founded in 1962, its mission was to provide its students an education concentrated on math, physics, and computer science.
I developed GeoDE for my final project at Carleton. Developed for use in learning and research, it provides an easy way to query, graph, and analyze the ~100m rows of temperature data collected through Carleton's geothermal power system. It allows the Geology department to study subterranean rivers and analyze the system's effects on them.
Pyodide is a package that lets you use Python alongside Javascript in the browser. I'm working on adding support for several Python packages, such as OpenCV and face-recognition, as well as dlib, a useful C++ library.
Unit Converter is a simple to use application that can converts between many units of measurement for length, volume, speed, etc., as well as between many currencies. It was developed by me and my classmate Dmytro Ivakhnenkov.