We are now building the next generation of Proof AI: systems that do not just surface the right information, but can take a compliance or sustainability requirement and produce a complete, evidence-backed, audit-ready result with expert oversight. This is some of the most ambitious applied-AI work at Sunhat, and it will shape where our platform goes over the coming years.
As an AI Product Engineer (m/f/d), you will help build this system, working close to where the value lands. You will see real customer assessments, iterate against live deliveries, and your work will directly determine whether a customer's compliance outcome ships on deadline. You will work as a full-stack engineer with a strong product mindset: building the agentic workflows and services (and the interfaces around them) that turn raw inputs into finished, audit-ready results, integrating and evaluating LLMs, and instrumenting the feedback loops that make the AI measurably better with every delivery.
Activities
- Drive generative and agentic AI from ideation to production, including multi-step agents that analyse, draft, verify, and prepare compliance deliverables (with human checkpoints where they matter).
- Own evaluation and quality: confidence scoring, accuracy measurement, and regression testing, so output is trustworthy enough to put in front of a customer.
- Instrument the data flywheel from day one, capturing each AI draft, expert edit, and final result so every delivery makes Proof AI measurably better.
- Build the product end to end, from the services and APIs to the lightweight interfaces the delivery team and customers use to review and approve work, plus the automation and integrations (e.g. via MCP) that connect them to the systems work flows through.
- Work forward-deployed, close to live deliveries: see where the product and the AI break under real deadlines, and turn that directly into what you build next.
- Partner with product management, compliance and sustainability experts, and the engineering team to translate delivery needs into reliable features, applying modern LLMOps practices and TDD for maintainable, rapidly iterated pipelines.
Requirements
- Proficient in GenAI and Agentic workflows: Multiple years of hands-on experience in applied AI such as generative models (Gemini, Mistral, Claude, Llama or similar) in production, particularly for text-based tasks. Experience with Vertex AI is a huge plus.
- Programming and Model Integration Skills: Expertise in TypeScript and experience integrating language models into production systems with a focus on reliability, scalability, and security. Expertise in Python is a plus but not a must.
- Fullstack Development Skills: Skilled in building server-side components in TypeScript or Python. You understand the full spectrum of client-server architecture. Ideally, you have worked with frameworks like Angular and Nest.js before.
- LLMOps and CI/CD Experience: Knowledge of CI/CD practices, ideally with GitHub, to streamline and automate model development and deployment.
- SaaS Experience: You have worked on B2B SaaS applications and leveraged AI.
- Data Security and Performance Optimization: Strong understanding of security, data privacy, and performance tuning specific to machine learning and GenAI in cloud environments.
- GenAI Expertise: Proficiency in applying GenAI to enhance user experiences, with a good grasp of advanced text generation, summarization, and other generative tasks relevant to sustainability data processing.
- Collaborative and Quality-Driven: Values collaboration, engages in pair programming, and prioritizes high-quality work through peer code reviews.
- Initiative and Adaptability: Takes the lead on new projects, drives them to production, and continuously learns and improves based on outcomes.
Team
We are a small, highly effective, and deeply product-minded team focused on delivering immediate value to sustainability and compliance teams. We believe in high autonomy, high responsibility, and building a foundation for scale.
- Development Culture: We are committed to daily shipping and leveraging automation wherever possible (CI/CD via GitHub). We prioritize code quality through peer reviews and dedicated testing practices (Testing Library, Testcontainers).
- Focus Areas: A key part of our work involves the development of our Proof AI (leveraging providers such as Gemini and Mistral) to make complex sustainability reporting effortless.
Application Process
- Discovery Call
- Live Coding Challenge with Engineer
- Technical Interview with 1-2 Engineers