Fullstack ML Engineer (LLM/ Langgraph) - Full-time - Worldwide

Company & Product

Finiti is building AI‑native software for SEC reporting. We are building an AI “corporate lawyer” that drafts, reviews and stress-tests corporate legal filings. Tech under the hood:

• First team to commercialise Stanford’s AI legal persona research into product (NeursIPS 2025); two patents pending
• Multi-agent orchestration with serial/parallel workflows
• Domain-specific prompting + context-aware drafting
• Near-infinite context approximation
• Not just “RAG++” - real architecture-level work

Role Overview

You’ll work closely with the founding team to turn legal logic into intuitive UX, orchestrate multi‑agent LLM systems for high‑stakes workflows, and ship reliable, production‑ready features quickly. We are targeting a launch on November 28, 2025.

What You’ll Do

• Architect and train domain-specific LLM systems with an emphasis on accuracy, explainability, and data provenance.
• Build and optimize retrieval-augmented generation (RAG) and graph-based context scaffolding for long-document reasoning.
• Orchestrate multi-agent pipelines (LangGraph / MCP / PydanticAI) with deterministic control and evaluation.
• Collaborate with full-stack engineers to integrate ML workflows into enterprise environments (Azure / Next.js / Python).
• Develop evaluation and benchmarking frameworks (precision, recall, F1, latency) to measure reliability across filings and workflows.

You Should Have

• Deep expertise in applied ML / NLP / LLMs (fine-tuning, retrieval systems, prompt orchestration, agent frameworks).
• Experience designing or training high-accuracy, domain-specific models — not just consuming APIs.
• Proven ability to handle complex data pipelines and retrieval tasks across unstructured corpora.
• Bonus: hands-on experience building or scaling multi-agent systems beyond surface-level orchestration.
• Bonus: strong full-stack or MLOps skills (React, FastAPI, Azure, Docker, LangGraph).

We would like you to be

• Self-starter, curious, “relentlessly resourceful” and fast-moving
• Cares more about building at the bleeding edge of applied LLM than chasing what’s hot
• Gets excited about novel business models and eating a $400B market from the inside