Особисті дані приховані

DANYLO ROSHCHENKO
PYTHON SOFTWARE ENGINEER
[відкрити контакти](див. вище в блоці «контактна інформація») [відкрити контакти](див. вище в блоці «контактна інформація») [відкрити контакти](див. вище в блоці «контактна інформація»)

Ukraine, Kyiv

PROFILE
Results-oriented Python Software Engineer with over 4,5 years of experience in designing and developing scalable
backend systems and LLM-powered AI solutions. Specialized in microservices architecture, complex payment
platforms, and building autonomous AI agents using LangChain and LangGraph.

Strong advocate of Clean Architecture, code quality, and system reliability. Passionate about taking ownership of
complex features from design to production, mentoring developers, and delivering measurable business impact
through technical excellence.

Education: Bachelor of Economic Cybernetics - National Aviation University - 2019 SEP - 2023 JUL
Language: English - B1 to B2

SKILLS
Languages: Python • JS/TS
Frameworks: FastAPI • Django/DRF/Django-channels • LangChain • LangGraph • Celery/Dramatiq
AI / GenAI: LLM Providers(OpenAI/Antropic/Gemini/Mistral/QWEN) • Agent System • Fine-tuning • LoRA/QLoRA •
vLLM/Ollama • HuggingFace • PyTorch • Transformers • GenAI • AI Assistants • RAG System • STT/TTS • n8n
Databases: SQL/NoSQL • PostgreSQL • MongoDB • DynamoDB • Vector Database• Elasticsearch • SQLAlchemy •
Alembic • Redis
Messaging & Async: RabbitMQ/SQS/Redis • NATS • WebSockets • gRPC
Cloud & DevOps: AWS/GCP · Docker • Kubernetes • DataDog • Sentry • Grafana • Prometheus • Terraform
Architecture: Microservices Architecture • Clean Architecture • DDD • SOLID • Product Engineering/Product Mindset
Testing: Pytest • Mypy

WORK EXPERIENCE
Python Software Engineer at Develux - 2025 MAR – PRESENT
Built and launched LLM-powered features that delivered +12% Conversion Rate and +5% ARPU in the first quarter after release.
Developed Proofs of Concept (PoCs) and internal prototypes using Node.js to validate product ideas, integrations, and architectural
approaches before full-scale implementation.
Built AI agent system prototypes and workflow automations using n8n for rapid experimentation, orchestration, and validation of
agent-based workflows.
Fine-tuned and optimized LLMs using QLoRA (Quantized Low-Rank Adaptation) techniques to improve model quality, domain
adaptation, and inference efficiency while reducing hardware and training costs.

Designed and implemented a flexible multi-product subscription billing system supporting multiple domains and isolated tenant
configurations, while integrating various payment gateways including Stripe, NMI/EMB, ensuring high reliability and fault tolerance of
payment flows.

Architected modular microservices using clean architecture principles, which reduced feature development time from 2 weeks to 3–
4 days and significantly lowered development effort across the team.

Developed and optimized autonomous AI agent systems powered by LLMs, resulting in 2x reduction in processing time and 2x
reduction in operational costs — this helped improved unit economics, significantly improved unit economics and helped achieve
profitability, and increase margin.

Mentored and onboarded new developers (newcomers), conducted code reviews, assisted with project adaptation and permissions
setup, contributing to faster team ramp-up and improved overall code quality.

Technologies: Django, DRF, FastAPI, Celery, Dramatiq, LangChain, Langraph, RabbitMQ, NATS, Microservices Architecture, Event-Driven
Design, Clean Architecture

WORK EXPERIENCE

Python Software Engineer at ByneSoft - 2024 SEP – 2025 MAR
Developed a comprehensive Open API service with interactive Playground that enables secure on-premises deployment for
enterprise clients, supporting integration with multiple LLM providers including OpenAI, Anthropic, Hugging Face, TogetherAI, Mistral,
and Gemini.

Designed and implemented AI agent systems using LangChain, featuring custom execution layers, advanced prompt engineering,
multi-step reasoning, and tool-calling capabilities to enable autonomous agent workflows.

Enhanced agent intelligence by integrating embedding solutions and vector search, significantly improving context understanding
and response quality.
Migrated and built robust backend infrastructure from AWS cloud to support secure on-premises deployments, using FastAPI,
AsyncIO, Guardrails, Celery + Redis, and Terraform for reliable, observable, and production-ready LLM agent platform.

Technologies: FastAPI, Langchain, GuardRails, AWS, SQS and Lamda, DynamoDB, MongoDB(Motor), Celery and Redis, AsyncIO,
OpenSearch, Terraform, MinIO.

Python Software Engineer at vidby - 2023 MAY – 2024 SEP
Maintained and significantly optimized existing system functionality, resulting in improved overall stability, performance, and
reliability of the platform.

Refactored large parts of legacy code and led the partial migration of the system toward a microservices architecture, which
enhanced scalability and maintainability.
Worked extensively with speech and language technologies: integrated and optimized multiple LLMs (Llama, GPT, Bart), various
Speech-to-Text services (Microsoft, Google, Sonix, Whisper), and multiple Text-to-Speech providers (Amazon, Microsoft, OpenAI,
Google, ElevenLabs).

Developed and improved several AI-powered features, enhancing the core capabilities of the speech-to-text and text-to-speech
platform.
Onboarded and mentored new developers (newcomers), helping them quickly adapt to the project and codebase.
Successfully released a project for a freelance marketplace, including full implementation of payment integration and connection to
the main project via POP (point-of-presence) communication.

Created stable integration between the freelance marketplace project and the main platform, enabling seamless data exchange
and request processing through API communication.
Actively reduced technical debt, improved code quality, and enhanced system scalability through continuous refactoring and
architectural improvements.

Technologies: Python, DRF, FastAPI, Celery, Dramatiq, Redis, RabbitMQ, PostgreSQL, MongoDB, GCP, Sentry, Django-Channels.

Python Software Engineer at LaSoft - 2022 JUN - 2023 APR
Worked in an outsourcing environment with international teams and clients, collaborating across different time zones and business
domains while participating in cross-functional communication, delivery planning, and technical discussions.

Designed and developed a backend service from scratch for integration with a Smart Home system, fully responsible for
architecture, business logic, and implementation.

Implemented core functionality including real-time communication, device management, and data processing pipelines.
Built a reliable asynchronous architecture using Celery for background tasks and Django Channels + RabbitMQ for real-time
features.

Technologies: Python, DRF, Celery, PostgreSQL, Redis, AWS, Django Channels, RabbitMQ.

Python Software Engineer at Volunteering - 2022 MAR - 2022 JUN
Developed a web scraping application for parsing and analyzing data from social media platforms.
Implemented data processing pipeline and built a Telegram bot using Aiogram to display parsed analytics and insights to users in
real time.

Technologies: Aiogram, Thron, Selenium, SQL Alchemy, PostgreSQL, Pandas.