Senior Machine Learning Engineer, New Verticals Agentic Foundations

DoorDash USA; Seattle (WA); Hybrid; USD 168,000 - 247,000; Full time

Benefits offered by this job

Premium healthcare

Wellness expense reimbursement

Paid parental leave

Job summary

DoorDash is seeking a Senior Machine Learning Engineer to advance agentic foundations for vertical agents in a Seattle-based hybrid role. You will lead eval-driven improvements for flagship agents, including the Ask grocery assistant, and expand to Merchants and Dashers with production-ready approaches.

You will design eval harnesses, refine prompts and tools, and run rigorous experiments including fine-tuning open-weight models and deploying small language models to enhance customer tasks and

Qualifications

• 3+ years of ML experience with hands-on work shipping LLM-based agents.
• Experience with post-training or fine-tuning open-weight models (SFT/RL) incl. small LMs.
• Strong NLP and ML background; Python proficiency with PyTorch or TensorFlow.
• MS or PhD in CS/Stats/Math or equivalent experience; collaborative and impact-driven.

Responsibilities

• Own continuous improvement of flagship agents, starting with the Ask Assistant, via an eval-driven loop to boost speed, cost, and reliability for customers.
• Extend eval-driven approaches to Merchants and Dashers, turning ideas into production-ready agents.
• Build an eval harness and optimization loop to identify failure modes, iterating on prompts, tools, skills, and context.
• Run experiments incl. fine-tuning open-weights models and deploying SLMs where suitable.
• Identify agentic opportunities for Merchants and Dashers from early exploration to production.
• Collaborate on agentic memory and post-training SLMs for steerable recommendations in production agents.
• Partner with leaders to define an ML-driven strategy for DoorDash’s grocery and retail delivery business.

Skills

LLM-based agents

Python

PyTorch

TensorFlow

NLP

Model fine-tuning

ML experience

Education

M.S./PhD in CS/Statistics/Math

Tools

SFT

RL

Job description

DoorDash is hiring a Senior Machine Learning Engineer to help build and improve agentic foundations for vertical agents.

Responsibilities

• Own continuous improvement of flagship agents, starting with the Ask Assistant, using an eval-driven loop to increase speed, cost-efficiency, and reliability for customers.
• Transfer the eval-driven approach to new agentic experiences for Merchants and Dashers, converting promising concepts into production-ready agents.
• Build and refine an eval harness and optimization loop to identify failure modes, then iterate on prompts, tools, skills, and context while measuring impact on real customer tasks.
• Run rigorous experiments, including fine-tuning open-weights models and deploying small language models (SLMs) to replace or augment larger LLMs where applicable.
• Identify agentic opportunities for Merchants and Dashers from early exploration through production delivery.
• Collaborate on agentic memory, agent-compatible product representations, and post-training SLMs for steerable generative recommendation, and help bring these capabilities into production agents.
• Partner with engineering, product, and business leaders to define an ML-driven strategy for DoorDash’s fast-growing grocery and retail delivery business.

Requirements

• 3+ years of industry machine learning experience, including hands-on experience building and shipping LLM-based agents (tool use, context management, prompting, guardrails) and improving them using evals and data.
• Experience with post-training or fine-tuning of open-weights models (e.g., SFT, preference optimization, or RL), ideally including small language models; ability to make quality, latency, and cost trade-offs.
• Strong foundation in NLP and machine learning; proficiency in Python with PyTorch or TensorFlow.
• M.S. or PhD in Computer Science, Statistics, Math, or another quantitative field, or equivalent practical experience; collaborative, growth-minded approach with a drive for measurable impact.

Technologies

• Python
• PyTorch
• TensorFlow
• SFT
• Preference optimization
• RL

Benefits

• Premium healthcare
• Wellness expense reimbursement
• Paid parental leave

About the Team

• Agentic Foundations, part of New Verticals ML, builds core AI/ML intelligence powering high-quality, scalable agents at DoorDash.
• Proves capabilities through flagship agent experiences, including the Ask grocery assistant.
• Operates a closed loop: new models, evals, and representations improve agents, while real-world usage surfaces gaps that guide next foundations work.

Role Details

• Location: Seattle, WA (hybrid)
• Compensation: USD 167,800 - 246,800 per year
• Reporting line: engineering lead on the New Verticals AI/ML team
• Work style: hybrid with some time in-office and some time remote