My work and results
Case studies
Companies I’ve worked with
Allure TestOps Cloud
- Increased trial-to-subscription conversion from 5.7% to 20.7% (3.6x): automated funnel reporting to show where users dropped off, then cut setup time from 12 to 3 minutes with faster DNS, and added a deployment-status page that made the wait feel shorter by showing each step and when the environment would be ready.
- Informed changes to pricing and usage limits by building an observability platform that connects tenant cost allocation with business metrics.
- Led the simplification of a platform serving approximately 900 tenants so one developer can manage it.
Kuvio
- Reduced the main analytics dataset’s cost from about $1,800 to $60 per month and a representative query’s runtime from 18.4 to 1.6 seconds.
- Identified wasteful data scans and implemented a fix that reduced the affected workload’s KMS and S3 costs from about $1,620 per month to near zero.
- Delivered while working one day a week on retainer.
- I also presented the S3 cost-optimization work at AWS Community Day.
Continue
- Automated delivery from GitHub across development, staging, and production, with Slack approvals giving the team control over releases.
- Built a reusable Terraform module for launching private or public ECS services with minimal configuration, built-in security controls, and automatic scaling.
- Added external uptime monitoring with automatic alerts for unavailable services and expiring certificates.
Sirob Technologies
- Built the RightStart AWS foundation later adopted by BORIS, separating workloads and shared services into different accounts.
- Configured private networking, encryption, and centralized access controls.
- Added audit logging, security monitoring, and budget alerts to support investigations and track spending.
Industrial Matrix
- Containerized three applications and prepared their AWS infrastructure so each service could be deployed and scaled separately.
- Built automated delivery across development, staging, and production.
- Helped developers adapt, debug, and deploy their applications on the new container platform.
- Deployed the application frontend on S3 and CloudFront.
My work and results
Allure TestOps Cloud
- Role:
- Lead AWS Consultant via FivexL
- Period:
- May 2025 – Present
- ⭐ Increased trial-to-subscription conversion from 5.7% to 20.7% (3.6x): automated funnel reporting to show where users dropped off, then cut setup time from 12 to 3 minutes with faster DNS, and added a deployment-status page that made the wait feel shorter by showing each step and when the environment would be ready.
- ⭐ Informed changes to pricing and usage limits by building an observability platform that connects tenant cost allocation with product usage and subscription revenue. It exposed customers whose infrastructure costs exceeded their payments, inactive environments still generating costs, and the unit costs of tests, attachments, and storage.
- Led the simplification of a platform serving approximately 900 tenants so one developer can manage it. Between June and August 2026, our team migrated tenant workloads from EKS to ECS and replaced self-managed messaging, caching, monitoring, and secrets infrastructure with managed AWS services. We retired the production EKS cluster on August 26.
Read more
- Supported Qameta’s completed SOC 2 Type II examination by implementing infrastructure security controls and preparing audit evidence. Restricted unnecessary public access and deployed temporary AWS access through SSO Elevator, replacing standing permissions for the support team.
- Built dedicated customer environments in separate AWS accounts, with private connectivity and a choice of hosting region. Automated deployment and removal, collected customer usage and cost data, and prepared cost estimates and architecture diagrams to help sales confirm customer requirements.
Kuvio
- Role:
- Senior Platform / DevOps Engineer via FivexL
- Period:
- Jul 2023 – Mar 2026
- Time commitment:
- Part-time, one day a week
- Case study:
FivexL
- ⭐ Redesigned Athena/S3 analytics using Apache Iceberg, partitioning, and file compaction. Reduced S3 spend by about $2,000 per month, reduced the main analytics dataset’s cost from about $1,800 to $60 per month, and cut a representative query’s runtime from 18.4 to 1.6 seconds. Presented this work at AWS Community Day.
- Traced 99% of 11.5M weekly KMS Decrypt events to a full-scan Athena pattern in a data-ingestion batch job and implemented the fix, reducing combined KMS + S3 spend from ~$1,620/month to near zero.
- Reduced AWS spend from about $9,500 to $7,700 per month by retiring old infrastructure after planning and testing the migration and rollback process. Removed unused databases, networking resources, servers, backups, and deployment pipelines.
Read more
- Evaluated a PostgreSQL-to-DynamoDB migration using the database schema, query patterns, data distribution, and estimated costs, and used DynamoDB Contributor Insights to identify throttled partitions while investigating reads taking up to one second.
- Reduced AWS governance, logging, and backup spend by tuning AWS Config, S3 access-log retention, and backup strategy; cut AWS Config from ~$157/month to ~$10/month, S3 access-log cost from ~$510/month to ~$112/month, and backup spend from ~$5.2K/month to ~$3.9K/month.
- Reduced CloudWatch custom-metric costs by about $390 per month by disabling unnecessary Glue job metrics. Identified an expensive DynamoDB-to-S3 export job; Glue costs fell by about $1,600 per month after it was removed.
- Protected the API’s main branch with maintainer approval and created a separate development environment for branch deployments, so developers could test changes before merging them.
- Traced database connection spikes to application connection pools and applied fixes to connection limits and cleanup. No further spikes were reported between April 24 and May 12, 2025. Later added RDS Proxy and a read replica to manage connections and distribute database load.
- Investigated missing orders and HTTP 502 errors by tracing requests across the load balancer, Envoy, ECS, and data-processing logs. Identified missing requests and data-format errors to guide production fixes.
- Standardized Terraform configuration and shared modules, brought unmanaged resources under Terraform, and upgraded deprecated dependencies. Migrated GitLab CI authentication from legacy JWT tokens to OIDC.
- Added AWS account security controls, including GuardDuty runtime monitoring, IAM Access Analyzer, service control policies, and updated TLS settings. Tuned AWS Config recording to control monitoring costs.
- ⭐ Built tools to collect Athena query data and analyze S3 access logs, showing which query and request patterns drove costs and guiding later Iceberg optimizations. Made AWS costs visible by resource using cost-allocation tags and S3 Storage Lens, and added budget and anomaly alerts to flag unexpected spending and guide cost investigations.
industrialmatrix.com
- Role:
- AWS Solutions Architect via FivexL
- Period:
- Jun 2025 – Aug 2025 · Support through Nov 2025
- Case study:
- FivexL
The client engagement included RightStart and ECS Blueprint. My contributions are described below; the published case study covers the broader FivexL engagement.
- Containerized three applications and deployed FivexL’s Blueprint for AWS ECS, including deployment pipelines and the required storage, queues, and database resources. Documented AWS access and service-health checks for the client team.
- Built automated delivery across development, staging, and production with separate settings for each environment. Fixed and deployed the frontend on S3 and CloudFront, and adapted the PHP/Nginx application for ECS.
- Helped developers adapt, debug, and deploy their applications on ECS. Guided the move from static AWS credentials to IAM-based authentication and configured VPN database access and ECS Exec.
FivexL: Blueprint for AWS ECS development
- Work:
- AWS Solutions Architect · Internal product development at FivexL
- Period:
- Mar 2025 – May 2025
- Built a Terraform-managed demo application on ECS/Fargate, including networking, service discovery, and routing between services.
- Added automated health checks and Better Stack monitoring to the demo application. Evaluated Playwright browser tests for session failures that basic HTTP checks could miss.
- Connected ECS metrics to Amazon Managed Service for Prometheus using AWS Distro for OpenTelemetry, providing monitoring data for the demo application.
FivexL: RightStart
- Work:
- AWS Solutions Architect · Internal product development at FivexL
- Period:
- Feb 2024 – May 2024
Worked on FivexL’s RightStart, a productized AWS foundation for startups.
- ⭐ Built and tested a prototype for automated AWS account setup using Control Tower Account Factory Customization, Service Catalog, and CloudFormation. Evaluated account-customization options and recommended Account Factory for Terraform for repeatable RightStart deployments.
- Built and tested security controls in a four-account AWS demo, preventing accounts from leaving the organization or disabling GuardDuty. Added Athena queries for reviewing S3 access logs.
- Automated cross-account network configuration and VPN setup by copying VPC and subnet metadata, replacing hardcoded VPC IDs, and automating certificate validation and SAML metadata sharing.
Allure TestOps Cloud
- Role:
- DevOps Engineer via FivexL
- Period:
- Jun 2021 – 2024
- Case study:
- FivexL
One of two FivexL consultants on the AWS rollout of Allure TestOps Cloud, the SaaS version of Qameta’s on-prem test management platform.
- Integrated Amazon Managed Grafana, Better Uptime, Prometheus, and CloudWatch monitoring and alerts during the second stage of the Allure TestOps Cloud rollout.
- Built an API for managing customer environments and maintained the Argo/Crossplane control plane, including recovery from failed resource deletions.
- Investigated EKS node failures, database-load spikes affecting 2–5 customer environments, and a Datadog agent running out of memory. Identified a faulty metrics collector during the database investigation and helped mitigate the incidents.
Read more
- Created an AWS infrastructure improvement roadmap covering deployment automation, access control, resilience across availability zones, cost allocation, commitment discounts, and storage cleanup.
- Helped Qameta’s new infrastructure lead understand AWS access management and the proposed RightStart setup.
Continue
- Role:
- AWS Solutions Architect via FivexL
- Period:
- February 2025
- Case study:
- FivexL
- Automated delivery from GitHub across development, staging, and production, with Slack approvals giving the team control over releases.
- Built a reusable Terraform module for launching private or public ECS services with minimal configuration, built-in security controls, and automatic scaling.
- Added external uptime monitoring with automatic alerts for unavailable services and expiring certificates.
Sirob Technologies
- Role:
- AWS Solutions Architect via FivexL
- Period:
- November 2025
- Case study:
- FivexL
- Built the RightStart AWS foundation later adopted by BORIS, separating workloads and shared services into different accounts.
- Configured private networking, encryption, and centralized access controls.
- Added audit logging, security monitoring, and budget alerts to support investigations and track spending.
More experience
Georgievsk Social Service Center
- Role:
- System Administrator
- Period:
- Jul 2018 – Jul 2021
- Time commitment:
- Full-time
- Maintained systems, installed software updates, and helped staff resolve technical problems.
- Helped introduce new systems and train staff, while following data-security and compliance requirements.
- In my free time, I learned about data scraping, cleaning, and visualization. You can find more on X.
Certifications
Verified by Credly
Open source contributions
terraform-aws-sso-elevator
64 stars
Terraform-based solution for just-in-time privilege escalation in AWS via SSO and a Slack bot, eliminating standing access by granting temporary permissions on demand and aligning with least-privilege security practices.
My contribution: Built the foundational architecture and access control logic.terraform-aws-s3-access-logs-athena-table
2 stars
Terraform module that creates an AWS Glue table designed specifically for querying Amazon S3 server access logs in the date-based partitioning format (optimizes query price and performance).
My contribution: Fully implemented by me.aws-eni-identifier
9 stars
Identifies which AWS service a network interface is associated with.
My contribution: Fully implemented by me.terraform-aws-ecs-events-to-slack
12 stars
Amazon EventBridge rules that fetch ECS events and send them to Slack.
My contribution: Co-authored configurable CloudWatch log retention, keeping the 14-day default.
Public activity
AWS Community Day — Central Asia · May 26, 2024
Title: Athena Best Practices and Beyond: The story of S3 bucket cost
Description: I presented how we reduced one S3 bucket’s costs by 96%. The talk covers tracing expensive Athena queries, understanding S3 request charges, and testing partitioning and Apache Iceberg file compaction to reduce costs. View presentation →