- Company Name
- Faire
- Job Title
- Applied AI / ML Scientist - Search Ads
- Job Description
-
**Job title:** Applied AI / ML Scientist - Search Ads
**Role Summary:**
Lead the design, development, and production deployment of machine‑learning models that drive relevance and personalization for Search Ads in a real‑time advertising platform. Own the entire relevance stack—from query understanding to final ad placement—while ensuring strict latency, scalability, and reliability targets.
**Expectations:**
- Deliver production‑grade ML models that enhance sponsored‑result relevance and advertiser ROI.
- Translate research into measurable impact through rigorous offline evaluation and online experimentation (A/B).
- Collaborate closely with software engineering, data engineering, and product teams to integrate models into the broader ads marketplace.
- Shape technical roadmap and culture for a nascent Ads Data team, leveraging product intuition to define effective advertising solutions.
- Maintain high standards of code quality, documentation, and testing for reproducible releases.
**Key Responsibilities:**
1. Own and evolve the Search Ads relevance architecture → query understanding, candidate generation, multi‑stage ranking, calibration.
2. Design, train, and productionize deep‑learning models that combine LLM‑based representations (BERT, GPT‑style) with structured marketplace features.
3. Optimize models for low‑latency inference and high throughput, meeting real‑time constraints in ad auctions, budget pacing, and pacing constraints.
4. Conduct offline analysis (log‑based metrics, user embeddings, calibration) and online experiments to link relevance improvements to business outcomes (CTR, conversion, long‑term retailer satisfaction).
5. Define and refine evaluation pipelines, feature stores, and embedding indexes (Faiss, ScaNN, Pinecone).
6. Engage in cross‑functional planning sessions to prioritize experiments, toolchains, and feature flags.
7. Mentor peers, establish best practices for reproducibility, unit testing, and continuous integration.
8. Stay current with cutting‑edge research in NLP, LLMs, and recommendation systems, and rapidly prototype promising techniques.
9. Communicate model performance, trade‑offs, and business impact to stakeholders at all levels.
**Required Skills:**
- 2+ years of building and shipping production ML systems in search, recommendation, or ad‑ranking domains.
- Proficiency in PyTorch (or equivalent DL framework) and vector‑search libraries: Faiss, ScaNN, Pinecone.
- Demonstrated experience integrating LLM embeddings with structured features to drive relevance and personalization.
- Strong product mindset: translating model insights into tangible user/business outcomes.
- Solid programming skills in Python; familiarity with distributed training/inference pipelines.
- Ability to design, implement, and optimize low‑latency inference services (e.g., ONNX, TorchScript).
- Excellent verbal and written communication; adept at collaborating in interdisciplinary teams.
- Empirical evaluation skills: offline metrics, online A/B testing, statistical significance.
**Required Education & Certifications:**
- Bachelor’s degree in Computer Science, Data Science, Statistics, Electrical Engineering, or related STEM field (mandatory).
- Master’s or PhD in Computer Science, Statistics, Machine Learning, or a closely related discipline highly preferred.
**Certifications:**
- Not required, but certifications in deep learning (e.g., TensorFlow, PyTorch) or data engineering are a plus.
San francisco, United states
Hybrid
Junior
18-03-2026