Skip to content

Hi, I’m

KaterynaOblakevych

Senior Software Engineer

Distributed Backend Systems · Cloud Platforms
AI-Powered Workflows

View ExperienceNotes
An engineer waters a laptop on a small table while flowers grow out of it — a metaphor for backend systems that grow stronger with care.
  • Observability
  • APIs
  • Reliability
  • AI Workflows

Tracing · Data Integrity · Security · Events

The principle

Systems don’t become reliable by accident.

Good systems grow from thoughtful architecture, clear contracts, observability, and care in production.

I think of backend systems as living ecosystems. APIs, data, infrastructure, security, and the people around them all have to work together.

My work focuses on building systems that don’t just function, but stay healthy as they scale.

Selected Experience

Current

Microsoft

Senior Software Engineer

Reliable backend foundations for AI-assisted investigation workflows.

  • Integrated backend services with LLM/VLM document-analysis pipelines.
  • Built validation layers and API contracts for model-generated outputs.
  • Improved tracing, diagnostics, and pipeline visibility for faster investigation.

LLM/VLM · Validation · Tracing · Distributed Services

Crowdin

Earlier

Senior Software Engineer

Backend services and event-driven pipelines for localization workflows at scale.

  • Built file-processing and customer-facing APIs for localization workflows.
  • Designed event-driven processing with object storage and serverless compute.
  • Improved API latency, authentication consistency, and deployment reliability.

Event-Driven · Object Storage · Serverless · API Performance

Sigma Software

Earlier

Software Engineer

Java/Spring Boot microservices with authentication, testing, and reliability foundations.

  • Built production microservices, backend APIs, and asynchronous processing flows.
  • Expanded automated testing and JWT-based service authentication.
  • Improved scalability, reliability, and reduced production defects.

Spring Boot · JWT · Testing · Microservices

Public Work

/realtime-rate

Realtime Rate

Rate-limiting and streaming-rates backend experiment.

GoWebSocketsRedis
/machine-age-validator

Machine Age Validator

Validation service for strict input contracts and defensive request handling.

PythonFastAPIValidation
/uni-manage

Uni Manage

University management backend with CRUD APIs and relational data modeling.

JavaSpring BootPostgreSQL

Skills & Systems

Build

  • Java
  • Go
  • Python
  • TypeScript
  • Spring Boot
  • REST APIs
  • GraphQL
  • Microservices
  • Event-Driven Architecture
  • Kafka
  • RabbitMQ

Run

  • AWS
  • Azure
  • Docker
  • Kubernetes
  • ECS
  • Serverless
  • CI/CD

Store

  • PostgreSQL
  • Redis
  • DynamoDB
  • S3-Compatible Storage
  • Flyway
  • Liquibase

Observe

  • Prometheus
  • Grafana
  • Elasticsearch
  • Kibana
  • Structured Logging
  • Tracing
  • Incident Triage

Protect

  • OAuth2
  • JWT
  • API Security
  • Vault
  • OWASP Top 10

Extend

  • LLM/VLM Integrations
  • RAG
  • Model-Output Validation
  • Spec-Driven Development

Notes

01

Observability for AI-assisted workflows

Observability

What you actually need to see when a model sits in the middle of your pipeline.

02

Why validation layers matter in LLM-powered systems

Validation

Treating model output as untrusted input — and the contracts that make it safe.

03

Designing backend systems that stay healthy in production

Architecture

The habits that keep services predictable long after the launch adrenaline fades.