AI Engineer

Farsight AI•New York City, NY

10d•Hybrid

About The Position

As an AI Engineer at Farsight AI, you'll play a critical role in advancing our state-of-the-art AI platform. You will: Design, build, and own agentic systems that produce client-ready financial deliverables — decks, models, and memos — end to end, from data ingestion to polished output. Build the evaluation and quality systems behind generated deliverables: deterministic document checks, model-judged review, grounding and fidelity verification, and regression suites for non-deterministic outputs. Turn proprietary data from our platform and customers into evidence-backed insight — building pipelines that discover, quantify, and verify patterns, and feed them back into how our systems draft and check. Build and refine the retrieval systems that ground our deliverables in unstructured and structured financial data, ensuring every number and claim traces to its source. Drive the next generation of agentic capabilities on the Farsight platform — composability, advanced function calling, and modular orchestration. Shape what we build next: ideate directly on customer value, and help prioritize across time-to-market, execution risk, and impact.

Requirements

2+ years of experience deploying machine learning or LLM systems in production.
Experience building agentic systems: function calling, tool use, orchestration, and the reliability engineering that keeps multi-step systems on the rails.
Experience building and validating evaluations for LLM systems.
Experience building retrieval over unstructured and structured data, and working with documents as structured artifacts (mixed text, numeric, and tabular content; document formats and their internals; rendered-output analysis).
A real, evolving AI-assisted development workflow, with a verification story proportional to what you delegate.
Strong Python proficiency and deep, current fluency with frontier model APIs and their capabilities.

Nice To Haves

Experience evaluating generated documents or other complex, client-facing deliverables — especially combining deterministic checks with model-based judging.
Experience extracting reliable, quantified patterns from large, noisy real-world datasets — discovery pipelines where every claimed finding traces back to specific evidence.
Deep familiarity with financial data, documents, and workflows.

Responsibilities

Design, build, and own agentic systems that produce client-ready financial deliverables — decks, models, and memos — end to end, from data ingestion to polished output.
Build the evaluation and quality systems behind generated deliverables: deterministic document checks, model-judged review, grounding and fidelity verification, and regression suites for non-deterministic outputs.
Turn proprietary data from our platform and customers into evidence-backed insight — building pipelines that discover, quantify, and verify patterns, and feed them back into how our systems draft and check.
Build and refine the retrieval systems that ground our deliverables in unstructured and structured financial data, ensuring every number and claim traces to its source.
Drive the next generation of agentic capabilities on the Farsight platform — composability, advanced function calling, and modular orchestration.
Shape what we build next: ideate directly on customer value, and help prioritize across time-to-market, execution risk, and impact.