Full Stack AI Engineer
Related Skills & Technologies
Professionals who know software development often also have expertise in:
The Story Behind the Role
Given the rapid pace of development and high autonomy, how does Zalion ensure individual contributors feel supported in their growth and prevent burnout while pushing for aggressive goals?
That's a fantastic question, and it gets right to the heart of how we operate here at Zalion. While it's true we move incredibly fast and give our engineers a huge amount of ownership, we're not about burning people out. Our support system is baked into our 'get stuff done' mentality. We foster a culture where collaboration is key – you're never truly on your own when tackling a complex problem. We encourage sharing knowledge, peer reviews are thorough yet constructive, and leadership is always accessible for guidance, not just direction. Plus, because we value impact over hours, we trust you to manage your own time effectively, ensuring you have the space to recharge. We believe that true growth comes from tackling hard problems with the right support, not just working longer.
The job mentions working with both structured and unstructured data, often integrating with 'real-world messiness' in legacy systems. Can you elaborate on the most significant technical challenges you've faced in building robust AI agents for such varied data sources and how Zalion approaches mitigating these complexities?
That's a very keen observation, and it highlights one of the most exciting, yet challenging, aspects of our work. Integrating with real-world enterprise systems is rarely a clean process. We often deal with incomplete data, inconsistent formats, archaic APIs, and systems that weren't designed for AI agents to interact with them. Our approach is multi-faceted: first, we invest heavily in robust data validation and cleaning pipelines, treating data quality as a first-class citizen. Second, we design our agents with extreme resilience and error handling, anticipating failure modes and building sophisticated recovery mechanisms. We leverage tools like LangGraph to orchestrate complex flows, making them debuggable and observable. Finally, and crucially, we embrace an iterative, 'builder mindset'. We start with simpler interactions, learn from the 'messiness,' and incrementally refine our agents, often collaborating directly with client IT teams to understand the intricacies of their systems. It's about engineering pragmatism and a commitment to making agents trustworthy, not just 'magic'.
Description
Zalion is an international AI start-up revolutionizing industrial procurement by using cutting-edge AI agents to automate manual Source-to-Pay processes. This frees up procurement teams for more strategic work.
The company is already live with some of the largest companies in Europe and the US, demonstrating tangible impact. Zalion fosters a culture of passion, efficiency, and rapid progress, where employees "love what we do," "get stuff done," and "progress fast."
What You'll Do
- Own and ship agent-powered features end-to-end, from initial idea and prototype through to production rollout and iteration.
- Design, build, and deploy AI agents into production environments that interact with real systems such as ERPs, email platforms, and supplier portals, delivering measurable outcomes.
- Work effectively with both structured and unstructured data, including RFQs, emails, quotes, and contracts, to power agent memory, retrieval mechanisms, and decision-making processes.
- Build scalable memory, reasoning, and orchestration pipelines, leveraging tools like LangChain/LangGraph and robust evaluation frameworks.
- Integrate with external APIs and legacy systems, navigating real-world complexities such as data messiness, rate limits, and various failure modes.
- Contribute to the development and maintenance of Zalion's cloud infrastructure to support scalable agentic workloads, focusing on observability, CI/CD practices, reliability, and cost-awareness.
- Solve complex problems related to context management, latency, robustness, and guardrails, ensuring AI agents are debuggable and trustworthy, rather than opaque.
What You'll Need
- Strong proficiency in Python and a proven track record of writing clean, maintainable, and well-tested backend code.
- Solid experience in designing and building RESTful APIs and data models for production environments.
- Experience with key data engineering concepts, including ETL processes, data quality assurance, and performance optimization, alongside strong backend architecture skills.
- Hands-on experience building a real application or system with Large Language Models (LLMs) from providers like OpenAI, Claude, etc. (concrete side projects are welcome).
- Comfort working in cloud environments (ideally AWS) with practical experience in CI/CD pipelines, logging, and monitoring.
- Possess a true \"builder mindset\" – a preference for shipping small, iterating quickly, and taking full end-to-end ownership of your work.
- Comfortable with ambiguity and demonstrating strong ownership: you proactively identify problems, propose solutions, and drive them to completion without requiring perfect specifications.
- Roughly 4+ years of professional experience, including at least one role where you were solely responsible for a production service, subsystem, or a major feature.
- **Python**: LangChain, LangGraph, LangSmith, FastAPI
- **AWS**: Bedrock, ECS Fargate, Lambda, S3, SQS/SNS
- **Frontend**: Next.js, React, TanStack Query/Router
- **DevOps/Monitoring**: GitHub CI/CD, Sentry, Grafana, Prometheus
- **Databases**: Postgres, Redis
- **AI Infrastructure**: Large Language Models (LLMs), evaluation frameworks, tracing, guardrails
Our Culture
Provenance Trail
To ensure authenticity, we track the origin and verification history of our listings.
- Source:arbeitnow
- Original Posting:View Original