Principal Data Scientist, Agentic AI Technical Lead

robotsandpencils; United States; Remote; USD 250,000 - 420,000; Full time

Job summary

robotsandpencils is seeking a senior data scientist and AI engineering leader to act as technical lead for a large agentic AI program. You will shape architecture, evaluation, and delivery while advising executives across multiple streams.

The role calls for setting direction, driving quality, and owning critical decisions when stakes are high. You will define strategies for multi-agent orchestration, tool calling, RAG, and streaming responses, while ensuring robust evaluation and governance

Qualifications

• 12+ years in data science and AI/ML with production-scale AI systems.
• Ability to communicate credibly with senior stakeholders and client leadership.
• Hands-on with LLMs, prompt engineering, multi-agent orchestration, RAG, and vector databases.
• Deep expertise in model evaluation, experimentation design, and statistics.
• Strong Python skills and fluency with the ML stack.
• Background in MLOps, model versioning, monitoring, and reproducibility.

Responsibilities

• Provide senior technical oversight across all agentic AI workstreams from framing to deployment.
• Define technical strategy and reference architecture for agentic systems and tools.
• Manage cross-team dependencies, remove blockers, and sustain velocity and quality.
• Set standards for model development, experiments, and production transition.
• Establish evaluation frameworks covering quality, safety, cost, and latency.
• Mentor technical leads and own high-stakes decisions across the program.

Skills

Python

Executive presence

Stakeholder communication

LLMs & agentic systems

MLOps

Evaluation design

Tools

Vector databases

Embeddings

Model monitoring

Deployment automation

Job description

Role overview

A senior leadership seat for a deeply experienced data scientist and AI engineer to act as technical lead across an entire agentic AI program. The role spans architecture, evaluation, delivery management, and executive advisory across multiple workstreams inside a large delivery organization. It is built for someone who sets direction, raises quality, and owns the call when the stakes are high.

Responsibilities

• Provide senior technical oversight across every agentic AI workstream, from problem framing through evaluation, deployment, and ongoing operation
• Define technical strategy and reference architecture for agentic systems, including multi-agent orchestration, tool and function calling, RAG, vector databases, embeddings, and streaming responses
• Manage cross-team dependencies, resolve blockers, and keep velocity and quality high as the program scales
• Set standards for model development, experimentation, and the path from research to production
• Establish evaluation frameworks covering quality, safety, hallucination, cost, and latency for LLM and agent systems
• Mentor technical leads, champion engineering standards, and own high-stakes decisions across the program

Requirements

• 12+ years of experience in data science and AI/ML with a record of taking AI systems from research to production
• Strong executive presence and the ability to communicate credibly with senior stakeholders and client leadership
• Hands-on experience with LLMs and agentic systems, including prompt engineering, function and tool calling, multi-agent orchestration, RAG architectures, vector databases, embeddings, and streaming LLM responses
• Deep expertise in model evaluation, experimentation design, and applied statistics, including evaluation approaches specific to generative and agentic systems
• Strong Python skills and fluency with the modern data science and ML stack
• Background in MLOps and AI infrastructure, including model versioning, monitoring, deployment automation, and reproducibility

Nice to have

• Experience supporting AI programs in regulated industries such as life sciences, healthcare, or financial services
• Consulting or client-facing technical leadership experience
• Background in fine-tuning, evaluation automation, or agent safety and guardrails