SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 200,000 - 300,000 / annual
Farsight AI is an agentic AI platform for financial services, helping investment banks and private equity firms automate complex, nuanced workflows. The company addresses a critical pain point: institutions currently spend 80% of their time gathering information and creating first drafts of materials manually before iteration and refinement. Farsight's AI platform returns that time to professionals while surfacing new insights they couldn't uncover independently.
As an AI Engineer, you will design, build, and own end-to-end agentic systems that produce client-ready financial deliverables including decks, models, and memos. Your responsibilities span the full pipeline: data ingestion, processing, and polished output generation.
Key responsibilities include:
- Building evaluation and quality systems for generated deliverables, combining deterministic document checks, model-based review, grounding verification, and regression suites for non-deterministic outputs
- Transforming proprietary platform and customer data into evidence-backed insights through discovery pipelines that identify, quantify, and verify patterns
- Developing and refining retrieval systems that ground deliverables in unstructured and structured financial data, ensuring traceability of all numbers and claims
- Advancing agentic capabilities including composability, advanced function calling, and modular orchestration
- Collaborating directly on customer value and helping prioritize across time-to-market, execution risk, and impact
Required qualifications include 2+ years deploying ML or LLM systems in production, hands-on experience building agentic systems (function calling, tool use, orchestration), expertise in LLM evaluation methodology, experience with retrieval over mixed data types and document formats, strong Python proficiency, and deep familiarity with frontier model APIs.
Desirable experience includes evaluating generated documents and complex deliverables, extracting reliable patterns from large noisy datasets, and deep knowledge of financial data, documents, and workflows.
The team includes founders and engineers from Evercore, General Atlantic, Menlo Ventures, Amazon, Hugging Face, and MIT Media Lab. The company is entirely based in Manhattan with in-person work 4 days per week (remote on Fridays).
