Full-Stack Engineer
Summary
Builds and deploys full-stack features in Python/Django APIs and React frontends, focusing on async workflows, AutoML inference, and AI-assisted development tools while ensuring clean, testable code and secure workflows.
About The Employer
Bezep Solutions is an AI and software development company building products for clients across Europe and the US.
Full-Stack Engineer
Location: Prishtina
Department: Engineering
Job Type: Hybrid, Full Time
Responsibilities
Job Description
- Implement and ship features across Python/Django APIs and React frontends
- Own vertical slices with guidance: models → serializers/views → services → UI → tests
- Work with async systems (Celery, Redis, SQS) for processors, predictions, and background jobs
- Extend AutoML / inference flows (predictions, converters, custom LLM-style paths)
- Write automated tests (pytest, frontend unit/Storybook tests; Playwright is a plus)
- Use agentic development tools (e.g. Cursor, Claude Code, or similar) to explore code, draft changes, and iterate—while keeping output reviewable and maintainable
- Help the team improve AI-assisted workflows: prompts, rules/skills, and safe day-to-day use of agents
- Collaborate with product and design; leave code cleaner than you found it
Requirements
- 3-5 years professional software experience (or equivalent depth)
- Strong Python fundamentals and willingness to grow deep in Django/DRF
- Solid React experience; can ship UI with light guidance
- Comfortable with REST APIs, relational databases, and learning async/background job systems
- Writes tests and cares about correctness
- Hands-on agentic development: you regularly use coding agents / AI IDEs to plan, implement, and iterate—and you know when not to trust the model
- Clear communicator; can work in small, reviewable PRs
- Comfortable with a hybrid work setup (in-office collaboration + remote focus time) Strong plus
- Experience with labeling/annotation, computer vision, or LLM product features
- AWS in production (S3, queues, auth)
- Celery/SQS or similar job systems
- Building agent loops yourself: tool calling, orchestration, evals, RAG, or multi-agent workflows
- Familiarity with Cursor rules/skills or similar repo-level AI tooling
- TypeScript discipline, Storybook-driven UI, or complex client state
- Security mindset around secrets, PII, and model/tool permissions