Founding ML Engineer

About the Role
We're looking for a Founding ML Engineer to build the systems that let us specialize models for real enterprise workflows — ServiceNow operations, KYB and client onboarding, marketing legal review, and due diligence document review. You'll work close to the data and close to the customer, turning messy enterprise inputs into fine-tuned, production-grade model behavior. This is a hands-on builder role, not a research role — you'll ship, measure, and iterate fast.

Key Responsibilities

- 🔨 Model Specialization : Fine-tune and evaluate LLMs against real customer workflows; build the eval harnesses that tell us when a model is actually ready for production.
- 🧱 Pipeline Ownership : Build and maintain the data pipelines — collection, labeling, cleaning — that feed model training and RAG retrieval.
- 🔌 Production Integration : Ship ML features directly into our product surface (Postgres, AI SDK, Northflank or similar), working alongside full-stack engineers.
- 📊 Experimentation : Run structured experiments on prompting, retrieval, and fine-tuning approaches; report clearly on what worked and why.
- 🤝 Cross-functional Work : Partner with product and the founding team to translate a vertical's messy real-world documents into a scoped, shippable ML problem.

Qualifications

- 3+ years of professional ML engineering experience, ideally with at least one full production ML system shipped end-to-end
- Hands-on experience fine-tuning LLMs and building RAG pipelines
- Comfortable working in TypeScript-based backends (Bun or Node.js) — you don't need to be a full-stack engineer, but you need to be able to read and modify the surrounding codebase
- Experience with messy, real-world enterprise data (not just clean benchmark datasets)
- Entrepreneurial mindset — comfortable with ambiguity and early-stage pace

Nice-to-Have

- Experience with managed LoRA/fine-tuning platforms (Together, Fireworks)
- Background in a AI-application vertical (legal, KYB/KYC, audit, customer support)
- Prior startup (0→1) experience