Talent.com
Principal GenAI Data Scientist

Principal GenAI Data Scientist

ZeniusMaclean, VA, United States
job_description.job_card.1_day_ago
serp_jobs.job_preview.job_type
  • serp_jobs.job_card.full_time
  • serp_jobs.filters_job_card.quick_apply
job_description.job_card.job_description

Principal GenAI Data Scientist

Duration : 3 months Start : ASAP Location : Onsite

Description

We are seeking a highly experienced Principal GenAI Data Scientist to lead the design and development of AI agents, agentic workflows, and production GenAI applications that solve real business problems. You'll be a hands-on technical leader who partners with full-stack engineers, designers, product managers, and data engineers to ship secure, reliable, and scalable GenAI solutions.

Responsibilities

  • Architect & build AI agents, agentic workflows, and end-to-end GenAI apps for diverse enterprise use cases.
  • Develop, fine-tune, and evaluate LLMs (e.g., Claude / Anthropic, Azure OpenAI, and OSS) and select the right model per use case (cost / latency / quality).
  • Design & deploy RAG and Graph-RAG systems using vector stores and knowledge bases; implement semantic chunking, metadata enrichment, and privacy controls.
  • Implement embeddings pipelines and integrate with vector stores (e.g., AWS Knowledge Bases / Bedrock, Elastic, MongoDB Atlas).
  • Leverage MCP (Model Context Protocol) and A2A (agent-to-agent) patterns to compose multi-agent solutions.
  • Build notebooks & services in Python / Jupyter; use SageMaker and MLFlow / Kubeflow on EKS for training, tracking, and deployment.
  • Curate enterprise data via connectors; orchestrate multimodal ETL / ELT (PDF / audio / video) with Spark / PySpark.
  • Integrate GenAI capabilities into enterprise platforms via APIs and standardized GenAI patterns.
  • Establish evaluation & safety : define automatic evals, bias testing, guardrails, and deployment readiness criteria.
  • Collaborate cross-functionally with UI / microservices teams to deliver polished, production solutions and measurable business value.
  • Document & mentor : codify patterns, playbooks, and best practices for repeatable delivery.

Must-Have Qualifications

  • Hands-on ML to GenAI transition with demonstrated delivery of AI agents / agentic workflows.
  • Deep experience with RAG (documents to vectors, retrieval, synthesis) and Graph-RAG.
  • Strong Python (Jupyter) and modern ML stack (Transformers, LangChain / LlamaIndex or similar).
  • Practical use of MCP and A2A communication patterns in real workflows.
  • Cloud-native AI on AWS (SageMaker, Bedrock; MLFlow / Kubeflow on EKS).
  • Vector databases / knowledge bases (AWS Knowledge Bases / Bedrock, Elastic, MongoDB Atlas, etc.).
  • Proven prompt engineering, fine-tuning, evaluation frameworks, and guardrails / safety implementation.
  • Built and deployed GenAI apps to production (latency, cost, observability, rollback, safety).
  • Strong data engineering fundamentals : ingestion, chunking, enrichment, anonymization, and governance.

    • Required Experience

  • 10+ years in AI / ML with 3+ years focused on applied GenAI / LLM solutions.
  • Prior software engineering experience and ability to partner closely with full-stack teams.
  • GitHub repository link required for consideration (please include recent GenAI / agent work).

    • Preferred Qualifications

  • Publications or patents in AI / ML / LLM.
  • Experience with enterprise AI governance and ethical deployment.
  • CI / CD for MLOps and scalable inference APIs; observability and evaluation in production.
  • Experience designing business use cases from problem framing through measurable outcomes.

    • Nice to Have (Role-Aligned Extras)

  • Multi-modal models (text / image / audio / video) and tool-use / function-calling.
  • Knowledge graphs for Graph-RAG; retrieval policy and query planning.
  • GenAI architectural patterns (routing, orchestration, distillation, hybrid search).
    • serp_jobs.job_alerts.create_a_job

    Principal Data Scientist • Maclean, VA, United States