Forward-Deployed Engineer W/ Generative AI Exp
REMOTE
6+ Months
ABOUT THE ROLE:
We're looking for a Forward-Deployed Engineer to join our Strategic AI team at Cohn Reznick.
* Develop creative solutions which utilize LLMs at scale, while remaining cost effective (including model selection, token allocation budgeting, and AI governance).
* Build enterprise AI for accounting, audit and tax workflows: documents and financial statements review, evidence extraction, and document classification.
* Utilize document intelligence to extract information with clarity, precision and detail.
* Build RAG ingestion pipelines that extract structure from unstructured sources (PDFs, scans, forms, financial documents)
* Run evaluations: LLM-as-a-Judge, Direct Preference Optimization and human evaluation, to measure efficacy and improve accuracy of LLM generation
OUR STACK:
Azure, Python, React, Typescript, Jinja, Django, Postgres, pgvector.
REQUIREMENTS:
- Significant experience in prompt engineering, RAG, evaluation, and LLM failure modes (hallucination, grounding, context management)
- Full stack experience; UI / API / DB and Back End experience
- Claude Code and GitHub CoPilot coding experience (auto coding to specs)
- Strong experience in Python, AI and agentic frameworks (any of the following: LangGraph, Claude Agent SDK, OpenAI Agents SDK, Google ADK, etc.)
- Hands-on experience with vector databases and embedding-based retrieval (e.g., any of the following: Open Search, Azure AI Search, pgvector, Pinecone, Weaviate)
- Azure and Cloud Services experience (containers, APIs, CI/CD, monitoring)
- Excellent communication skills; you can explain technical tradeoffs to partners, executives, and practitioners alike
NICE-to-HAVE:
- Graph RAG and/or Doc Intelligence Experience
- Experience with CPA, tax, and/or audit-based applications
- Knowledge of accounting and/or tax procedures and standards (e.g., GAAP, GAAS, PCAOB, IFRS, IRC)
- Willingness to travel occasionally to customer or team sites