Required Skills & Experience
Bachelor's degree in Computer Science or a closely related field. 3+ years of experience developing and shipping AI/ML products. Hands-on experience designing and building RAG pipelines with LangGraph. Experience with multi-node graph orchestration, state management, and prompt engineering across graph nodes. Demonstrated experience chunking, embedding, and indexing multiple content types for retrieval. Proven experience evaluating LLM and retrieval output quality and using evaluation results to improve system performance.
Nice to Have Skills & Experience
Master's degree in Data Science, Statistics, Engineering, or a related discipline. Experience with hybrid retrieval and late-fusion strategies, including dense vector search, sparse/keyword search, re-ranking, or reciprocal rank fusion. Experience integrating LLM pipelines with external systems through Model Context Protocol (MCP) or similar tool-calling frameworks. Experience connecting AI applications to live enterprise platforms and structured business data. Experience building AI agents and production-grade agentic workflows.
Job Description
We are seeking a mid- to senior-level Data Scientist to design and build AI products that help drive value throughout the M&A integration lifecycle.
This role will focus on evolving an existing LangGraph-based AI pipeline that combines hybrid, late-fusion RAG with live data from an acquisition-tracking platform through Model Context Protocol (MCP). The Data Scientist will also develop new AI agents, workflows, and use cases as opportunities emerge across M&A data.
The ideal candidate brings strong hands-on experience building production-grade LLM and RAG solutions and is comfortable partnering with business stakeholders, AI product managers, and application engineers to translate complex business challenges into scalable AI systems.
Key Responsibilities
Design, build, and enhance AI/ML products and agentic workflows supporting M&A integration. Develop and optimize LangGraph-based RAG pipelines, including multi-node graph orchestration, state management, and prompt engineering. Build retrieval pipelines across diverse data types, including unstructured documents, PDFs, tables, and structured records. Integrate LLM applications with live enterprise data and external tools through MCP or similar tool-calling frameworks. Evaluate LLM and retrieval performance across relevance, faithfulness, groundedness, and overall response quality. Use evaluation results to iteratively improve prompts, retrieval strategies, indexing, and overall pipeline architecture. Partner closely with business SMEs, AI product managers, and engineers to take AI solutions from concept through production.