Lead Data Scientist; Agentic Solutions

About Rivian

Rivian is on a mission to keep the world adventurous forever. We build emissions‑free electric adventure vehicles for curious, courageous souls who love the outdoors and want to protect it for future generations.

Rivian's AI team sits within the Analytics function of the Customer organization and builds the next generation of intelligent, autonomous systems that connect marketing, sales, and market intelligence. As a Staff/Lead Data Scientist focused on Agentic Solutions, you will design and operationalize the cognitive architecture that powers Rivian's AI agents—building reasoning loops, retrieval systems, and evaluation frameworks that allow agents to act on live business data with accuracy and reliability.

• Agentic Reasoning Loops:
  Design and program multi‑step reasoning frameworks using orchestration frameworks such as Lang Chain, Llama Index, or Haystack.
• Context Engineering & Advanced RAG:
  Architect advanced retrieval structures, experiment with embedding models, dynamic chunking strategies, and token management to ensure agents receive high‑fidelity business context.
• Proactive System Reasoning:
  Build internal logic and evaluation criteria for “Watchdog” agents, enabling them to analyze live operational data streams, infer anomalies, and generate proactive insights.

• Small Language Model (SLM) Strategy:
  Evaluate, select, and adapt small language models (e.g., Phi, Mistral, Gemma) for domain‑specific agentic tasks where precision, latency, and cost efficiency outweigh raw model scale. Define the decision framework for when to use an SLM versus a frontier model based on task complexity, context requirements, and inference cost.
• Token Economics & Efficient Agent Design:
  Architect agents and orchestration frameworks around token efficiency as a first‑class design constraint—develop strategies for context compression, prompt caching, dynamic context windowing, and call minimization to ensure every inference call is purposeful and cost‑effective at scale.
• Domain‑Specific Model Adaptation:
  Execute parameter‑efficient fine‑tuning (e.g., LoRA/QLoRA) and model distillation on open‑source models to embed Rivian’s fulfillment jargon, vehicle logistics vocabulary, and internal business rules, prioritizing targeted SLM adaptation over large‑scale LLM retraining when feasible.
• Rigorous Evaluation Frameworks:
  Establish statistical, model‑driven, and human‑in‑the‑loop testing benchmarks to validate agent reasoning, track accuracy drift, and minimize hallucinations.
• Behavioral Prompt Engineering:
  Continuously iterate on complex system prompting, structured output formats (e.g., enforcing strict JSON/tool schemas), and cognitive guardrails to keep agent behavior deterministic and aligned.

• Bachelor’s degree in Computer Science, Statistics, Mathematics, Engineering, or a related quantitative field.
• Master’s or PhD preferred.
• 5+ years of experience in Machine Learning, Applied Data Science, or AI Engineering with a proven record of designing cognitive frameworks, RAG systems, or agentic workflows for business applications.

• Expertise with modern AI orchestration frameworks (Lang Chain, Llama Index, Haystack) and deep understanding of LLM APIs, prompt engineering, and agent memory state management.
• Hands‑on experience with small language models (e.g., Phi, Mistral, Gemma) for domain‑specific deployment and the ability to reason about trade‑offs between model size, latency, cost, and task accuracy.
• Experience with parameter‑efficient fine‑tuning (LoRA/QLoRA) and model distillation.
• Deep understanding of token economics and experience designing agent frameworks that incorporate context compression, prompt caching, dynamic windowing, and call‑minimization strategies.
• Expert‑level proficiency with Python and SQL; comfortable using Databricks (Spark, Delta Lake) and working with structured schemas, metadata, and vector indices.
• Strong statistical foundation with familiarity in analyzing user behavioral data or event streams (e.g., Snowplow) to help…