ML Architect – Search & Recs, Real-Time Inference

Wonder; New York (NY); On-site; USD 180,000 - 260,000; Full time

Benefits offered by this job

Equity

401(k)

Medical, dental, vision plans

Disability coverage

Paid time off

Vacation

Sick leave

Job summary

Wonder is seeking a senior ML architect to own the end-to-end ranking and recommendation stack in New York. You will lead cross-system design, integrate LLM-driven understanding and embedding-based retrieval, and drive experiments to balance conversion with profitability.

You will mentor engineers and work with product and data teams to shape scalable, observable systems. You will be responsible for elevating engineering standards, building reproducible pipelines, and communicating complex

Qualifications

• MS/PhD in a quantitative discipline or equivalent experience.
• 4-8+ years building and shipping machine learning systems.
• For staff level and above: 3+ years operating at senior scope across multiple teams or systems.
• Deep experience in recommendation systems, ranking, or information retrieval at scale with latency and cost constraints.
• Production DL experience with TensorFlow or PyTorch, including training, serving, and tuning on GPUs.
• Experience with Large Language Models and transformer-based architectures, including fine-tuning and embedding generation.
• Strong data engineering fundamentals: PySpark, Hive/SQL, Python data stack.
• Experimentation fluency: designing AB tests and guardrails.

Responsibilities

• Lead the end-to-end architecture of the ranking and recommendation stack, including candidate retrieval, multi-objective ranking, calibration, and an ensemble balancing conversions and profitability.
• Make cross-system design decisions that individual model owners cannot complete alone.
• Drive evolution of optimization objectives from short-term conversions toward long-term diner value with offline evaluation and online experimentation.
• Integrate state-of-the-art research into runtime systems for information retrieval and recommender systems (LLM-driven query understanding, embedding-based retrieval, sequential user representations).
• Evaluate rigorously what transfers to production traffic and discard non-transferable approaches.
• Raise engineering and operational standards across teams (model evaluation, reproducible training pipelines, safe deployment, SLOs, observability, technical debt).
• Partner with Product, Search Engineering, Ads, and Data Platform to shape roadmaps and ensure data and infra exist before model needs are imminent.
• Mentor senior and mid-level engineers, participate in hiring, and grow technical depth so critical systems do not depend on a single person.
• Translate technical trade-offs into business terms for product and execs; document rationale clearly.
• Continuously improve business performance through deep analysis and exploring innovations.

Skills

Ranking systems

Information retrieval

Leadership

Mentoring engineers

Communication to executives

Self-directed learning

Education

MS/PhD in quantitative discipline

Tools

TensorFlow

PyTorch

GPUs

PySpark

Hive/SQL

Python

AWS/SageMaker

Job description

Wonder is seeking a senior ML architect to own the end-to-end ranking and recommendation stack in New York. You will lead cross-system design, integrate LLM-driven understanding and embedding-based retrieval, and drive experiments to balance conversion with profitability.

You will mentor engineers and work with product and data teams to shape scalable, observable systems. You will be responsible for elevating engineering standards, building reproducible pipelines, and communicating complex