Aleksandr Kolmakov

● Successfully converted partially manual Google spreadsheet analytics to fully automated Superset dashboard with DBT core and Mage including analytics data modeling overhaul.
● Introduced cetralized monitoring with federate Prometheus and Grafana that removed resource based outages.
● Created realtime RAG pipeline for support telegram bot that reduced amount of support tickets by half.

• Created streaming pipeline using Dataflow that allowed face emotion recognition frame by frame and sped
up current batch workflow by roughly 200% per customer.
• Introduced Airflow + K8S to already working infrastructure to reduce data inconsistencies between
departments databases which also reduced average parasitic load of production by 20%.

• Grew AI SaaS project for energy optimization from foundation to successful launch.
• Cut costs of AWS by 60% by conducting cost analysis and detaching environment instances into a separate fleet.
• Decreased time of training evaluation by 30% via realtime ETL and visualization pipeline.

• Flask + SQLAlchemy backend infrastructure from scratch while providing 100% test coverage.
• Implemented the successful migration to Neo4j to support complex conditional multiple join queries in under 0.3 seconds.
• Used sentiment analysis to extract keywords and boost suggestion accuracy of posts by 40%.
• Mentored 6 people to confident middle developer level.

• Improved performance of XML export pipeline by 30% by using proper replication techniques.
• Worked on geographical routing algorithms to optimize shipping time by 15%.
• Developed flight booking integration for travel portal that increased monthly usage by 120%

• Developed real estate web scraping pipeline and full text search application to increase search speed and
accuracy by 20%.
• Developed set of tools for network security check to assess 80% of common vulnerabilities and placed second
on CCNA CTF with their direct usage.