AI Engineer – Rapid Prototyping & LLM Systems Confidential Rebuilding Backbone U.S

About the Company:

This fast-scaling AI startup is redefining the future of tax and compliance a $70B industry under severe strain due to a critical talent shortage. Backed by leading investors and already generating significant recurring revenue, the company has achieved strong product-market fit in under a year. They’re not building another tool they’re creating an AI-powered teammate trusted by accounting firms nationwide.

This role is ideal for a highly pragmatic, technically sharp engineer who thrives in fast-paced environments. The team moves quickly—shipping to users within hours, not quarters. Engineers are expected to think in systems, make informed judgment calls amidst ambiguity, and own their outcomes from end to end.

This is not a research-for-research’s-sake role. It’s about shipping real products, solving messy real-world problems, and making an immediate impact on a sector in urgent need of innovation.

• Design and implement AI-powered workflows across tax, finance, and compliance use cases

• Build and scale structured memory and retrieval systems to support contextual AI

• Own full-stack ML components including prompt pipelines, vector stores, fine-tuned models, and evaluations

• Develop deep domain understanding and model user behavior to anticipate needs

• Ship quickly, iterate based on real feedback, and operate autonomously

• Deep understanding of large language models and their inner workings

• Proficiency with APIs such as OpenAI, Claude, and Gemini

• Strong experience with Python and cloud platforms (GCP or Azure)

• Familiarity with embeddings, RAG architectures, structured retrieval, and inference at scale

• Bonus:
  Experience building internal tools for debugging or evaluating ML models

• Fast-moving and ruthlessly pragmatic—knows when to ship and when to iterate

• Thrives on ambiguity, edge cases, and unsolved problems

• Collaborative and empathetic—able to work closely with product, engineering, and operations

• Deeply curious—eager to learn the intricacies of the domain being automated

• Outcome-focused—wants their work to drive measurable impact for real users

The company operates on a transparent, level-based structure with six tiers, each reflecting increasing ownership and impact. Reviews and promotion opportunities occur every four months, based on contribution—not tenure. Compensation is above market, with fast-growing equity and meaningful autonomy from day one.

• Rare opportunity to join a company early in its trajectory, with momentum and funding

• High-impact work in a mission-critical sector

• Real ownership and direct access to users

• No bureaucracy—just a fast, focused, and user-obsessed team