Full Stack Engineer
Summary
Build and deploy production-grade agentic AI systems using Python/Java and React, orchestrating multi-agent workflows and retrieval pipelines while ensuring reliability and observability.
BGTS is a software and technology solutions company with over 1,800 professionals and 25+ years of experience. Through engineering expertise and industry insight, our international offices deliver tailored solutions, enabling clients worldwide to achieve their business goals with flexibility, speed, and impactful results.
About the Role
BGTS International is looking for a skilled and proactive Full Stack Developer to join our team. In this role, you will work closely with our Head of Software Engineering, enterprise architects, and global clients across various industries, as well as driving key internal AI projects.
This is a hybrid role requiring 3 days a week in the office. Our office is located on the İTÜ Ayazağa Campus.
Responsibilities
- Design and build multi-agent and agentic workflows using LangGraph or comparable orchestration frameworks (state machines, tool use, memory, retries, and human-in-the-loop approval gates).
- Develop robust production backend services in Python and/or Java, ensuring clean API design, sound data modeling, and comprehensive test coverage.
- Build intuitive front-end interfaces (React or similar) that clearly expose agent behavior, including execution traces, intermediate steps, confidence levels, and points of human intervention.
- Implement and optimize advanced retrieval pipelines (chunking, embeddings, vector stores, hybrid search) and actively evaluate what improves actual answer quality.
- Instrument everything. Set up evaluation harnesses, tracing, cost/latency monitoring, and regression testing for both prompts and agentic graphs.
- Deploy and operate these systems on cloud infrastructure using containers and modern CI/CD pipelines.
Requirements
- 5+ years of building production-grade software in Python and/or Java
- Demonstrable experience with agentic frameworks—such as LangGraph, LangChain, CrewAI, AutoGen, OpenAI/Anthropic SDKs, or MCP—including systems that successfully made it past the prototype stage into production.
- A deep understanding of GenAI failure modes (hallucination, prompt injection, tool misuse, runaway cost, non-determinism) and the engineering patterns used to mitigate them.
- Comfort with Docker, Git, CI/CD pipelines, and at least one major cloud provider (AWS, Azure, or GCP).
- Clear and effective communication skills in English (both written and spoken) for direct client collaboration.
Nice to Have
- Experience in financial services or other highly regulated domains.
- Hands-on experience with model evaluation and observability tools (LangSmith, Langfuse, Weights & Biases, etc.).
- Background in Kubernetes, Infrastructure as Code (IaC), or general platform engineering.