Hi, I’m Petar. I am a software developer based in London.
When I’m not working I am day-dreaming of travelling the world and owning a little wood workshop. Currently trying to learn Spanish. For the 5th time.
You can get in touch with me at pgyurov93 at gmail dot com
Royal Holloway, University of London (2012 - 2015)
BSc Theoretical Physics — 1st Class Honours (1:1)
Keen AI (10/2021 - Present)
I am working on a variety of different projects, from object detection to aid efforts against Ash dieback disease to segmentation models used in corrosion assessment in steelwork. Concurrently developing our KAI platform that pieces some of this data together.
The Big Wild (07/2021 - 02/02/2022)
Machine Learning Developer — C++, dlib, Python
I helped The Big Wild prototype an object detector and landmark detector for polar bears which is part of a bigger pipeline that will eventually be able to distinguish individual polar bears using face detection. The aim is to tackle the difficulties faced in polar bear tracking by providing a less invasive and less resource consuming alternative.
New Zealand Department of Conservation (12/2020 - 03/2021)
ML Software Developer Volunteer — Python, Tensorflow, ElectronJS, Svelte
Developed Megadetector GUI, an open source, AI-assisted solution for labeling and sorting camera trap images. The desktop application is an integral part of the data processing pipeline for conservation scientists working to save endangered species. Built using ElectronJS, Svelte, Python, Tensorflow.
Makely (08/2019 - 12/2020)
Founder, Full-Stack Developer — Python, Flask, AWS, ReactJS, PostgreSQL
Makely is a 3D printing platform that connects everyday people to 3D printing enthusiasts and hobbyists, allowing them to print their ideas with virtually zero knowledge of the tech. I’ve currently put the project on pause. If you are interested in being a co-founder please get in touch.
Splash Damage (08/2017 - 08/2019)
Backend Engineer — Python, Flask, Golang, C#, AWS, Docker, Kubernetes, PostgreSQL
Developed the backend for the PC game Dirty Bomb and an another unannounced project. The live-service nature of the games required the backend(s) to be flexible, with low overhead on new features and reliable scaling.
Improved the player sentiment analysis using a Naive Bayes classifier. Prototyped an analytics pipeline to handle high-throughput events for all of the company’s titles.
Praesto Consulting (02/2016 - 05/2017)
Configured BI pipelines and developed custom data visualisations for large businesses such as Schneider Electric, Britvic and Madison Square Garden. Worked with a variety of large datasets, conceptually ranging from financial evaluations to athlete performance data.
Echogroup (08/2014 - 09/2014)
Junior Software Developer — Python, Tkinter
Developed a bespoke spreadsheet management software package from the ground up to handle the company’s customer information and automate the task of mailing batch and individual journals. Includes standard features such as refined searching and amending data, as well as unique and personalised sorting and comparison algorithms, encapsulated in a sleek and easy to use GUI.