buildbuild is a construction-focused project management software mainly focusing on SMBs. It binds all aspects of on-site construction work in one place to help businesses achieve transparency and get real-time feedback on their project's budgets.I'm responsible for:- Establishing the company's technical landscape and product vision- Building and managing strong development team (15 people)- Architecting the core functionality of our product utilising go language and ensuring optimal performance- Engaging in strategic business development initiatives to drive company growthStack: go (gqlgen), python, react, nextjs, kubernetes, postgres, redis, rabbitmq

pik.pro is a marketplace for small construction subcontractors to find jobs at PIK's construction sites. It enables employer to schedule small portions of work to the subcontractors, ensuring quality and providing payment and invoicing functionality. More than 10k subcontractors were working via pik.pro software achieving greater work conditions.- Formed technology landscape- Described and implemented algorithm for forming and assigning jobs to subcontractors (based on graph theory)- Managed work of 20+ developers across different product teams- Described and implemented all IT-infrastructure based on microservices- Integrated GraphQL framework as unite gateway to access all inner microservices- Described and managed stress testing for whole system- Increased test coverage of every back-end service up to 90%- Contributed to back-end and front-end servicesStack: python (fastAPI, django), nodejs (apollo), react, kubernetes, gcp, postgres, redis, rabbitmq

Dbrain had two major products — crowdsourcing platform to annotate and manage data for CV & NLP and AI-powered data extraction from documents.- Managed work of 15 developers and machine learning engineers across different product divisions.- Developed accounting part of the software- Managed product and project management for both labelling service and OCR service for B2B sales- Perform A/B tests to determine most effective solutionsStack: go (gin), python, react, docker swarm, postgres, redis, aws, TensorFlow, PyTorch

Connectome developed a real-time quality estimation software for product lines based on in-house CV and ML technologies.- Managed work of 3 developers- Developed back-end services for different parts of the systemStack: python, TensorFlow, react, docker swarm

Developed back-end services for different projects using go language, mongo, docker

Developed backend system for obtaining data from different versions of wearable children bracelets using go language, mongo, docker.

Developed high-loaded node.js back-end for gaming video streaming site.