Vlasii – Python, FastAPI, React

Document automation SaaS for creating, sending and tracking quotes, proposals and contracts, with e-signature and AI-powered document processing.

• Designed and implemented core parts of the metadata extraction pipeline using durable execution with Temporal, processing millions of events daily with 99.99% workflow completion reliability.
• Led the architectural redesign of AI agents, decomposing a monolith into modular use-case workflows and cutting cycle time by 40% via independent releases and isolated AI evaluations. Authored the RFC and coordinated implementation across multiple teams.
• Led the onboarding of 3 engineers to a newly formed team by creating documentation, conducting technical training and providing mentorship, enabling them to ship independently within 1 month.

Freight-tech product automating load sourcing and back-office work for trucking carriers, including LLM-powered email parsing, voice automation, and streaming data pipelines into a data warehouse.

• Implemented LLM-based automation for email parsing, eliminating manual errors and reducing processing time by 78%, allowing support resources to be reallocated to higher-priority work.
• Implemented streaming data pipelines using AWS Glue, Kinesis and Airflow, enabling low-latency ingestion of mission-critical data into the data warehouse.
• Migrated scheduled background jobs from AWS EventBridge + Batch to ECS with Celery + Redis, using Terraform for IaC. Centralized infrastructure and code, reduced failed jobs by 42% and achieved 99.99% uptime with no loss of critical scheduled data.
• Improved frontend maintainability by introducing reusable components, hooks, consistent linting and formatting standards.
• Integrated React applications with backend APIs, implementing routing, authentication flows, validation and robust error handling.
• Built complex frontend workflows with React and TypeScript, handling asynchronous operations, loading states, validation and error recovery.

Web platform helping business clients plan, track and control their supply chain operations.

• Refactored request-intensive endpoints to use asynchronous processing instead of multithreading, enabling the system to handle up to 90% more requests with the same resources.
• Implemented a Distributed Lock Manager to mitigate multi-server race conditions that had remained unresolved for over a year, eliminating critical authentication errors.
• Integrated Datadog and StatsD metrics across the application, increasing observability and reducing the number of production bugs.
• Developed and maintained frontend features in Vue.js, building UI components and integrating them with backend REST APIs.

Full-stack solutions for clients from enterprise to startup across Europe, the USA and Australia, in financial services, retail, travel and hospitality, and healthcare.

• Designed and built RESTful APIs and backend services with Python, Django, Django REST Framework and Flask, deployed on AWS (EC2, S3, RDS, Lambda).
• Developed client-facing web applications and internal admin tools with React, TypeScript and JavaScript, covering the full feature lifecycle from requirements to production release.
• Implemented asynchronous task processing and integrations with third-party services using Celery and RabbitMQ, improving the reliability of data synchronisation between client systems.
• Set up CI/CD pipelines in Jenkins with SonarQube quality gates, shortening release cycles and raising code quality standards across projects.