The Company

They are building the first Asset Certification Utility for the mortgage industry. They don't just "summarize" text; our Agentic AI executes compliance reviews, extracting and validating data against a "Source of Truth" that banks and regulators trust.

We are hiring a Mid-to-Senior Data Scientist who understands the math behind the model, not just how to import a library.

The Role

You will design and train ML systems from first principles. You will own the architecture for Agentic workflows that reason over complex, unstructured documents (PDFs, loan files) and execute multi-step tasks.

What You Will Actually Do

* Build, Don't Just Call: Design and train models from scratch. You aren't just calling OpenAI’s API; you are optimizing the inference engine.

* Agentic AI & RAG: Build systems where LLMs use tools, plan workflows, and retrieve context via advanced RAG (Graph-RAG/Neo4j preferred).

* Production Grade: Ship code that runs in production. This is an engineering-heavy data science role. You will handle failure modes, data drift, and deployment.

The Stack

* Core: Python (Production quality), PyTorch/TensorFlow.

* AI/LLM: LangChain, LlamaIndex, Vector DBs, Embedding pipelines.

* Data: Unstructured (PDF/Text) & Graph Databases (Neo4j is a huge plus).

The Profile

* Mid-to-Senior: You have built and shipped ML systems to production.

* Deep Tech: You know the difference between accuracy and precision and can explain why a model failed.

* Builder: You are comfortable working in a startup environment—autonomous, execution-focused, and ready to wear multiple hats.

The Offer

* Base Pay: $140K - $150K per year + Bonus + Equity

* Location: 100% Remote

* Impact: Join at the pre-Series A stage and own the core AI logic of a platform trusted by major custodians and banks.