Staff Data Scientist - AI & Intelligence

Search Atlas•San Francisco, CA

1d•$220,000 - $300,000•Hybrid

About The Position

Turn Data Into Decisions. Build Models That Ship. Lead the Science. Location: San Francisco, CA (Hybrid) The Mission Search Atlas hit $32M ARR bootstrapped. No VC. No safety net. Just terabytes of search data, millions of crawl signals, and AI agents making autonomous decisions in real-time. Our data isn't a report. It's the intelligence layer of agentic marketing. When our AI decides how to optimize a Fortune 500's entire search presence, your model’s power that decision. When we predict algorithm changes before they happen, your research makes it possible. We need a Staff Data Scientist who builds production ML systems, not Jupyter notebooks. You'll own critical models, define our research agenda, and ship algorithms that directly impact revenue. This is high-stakes data science at scale. San Francisco. In-person energy. Ship-or-die velocity. What Winning Looks Like Month 1: You've shipped a model improvement that increases agent decision accuracy by 15%. You know our ClickHouse schemas, feature pipelines, and model serving infrastructure cold. Month 3: You've led a major research initiative - ranking prediction, content optimization algorithms, or anomaly detection at scale. It's in production, not a slide deck. Month 6: You're the scientific authority for our AI systems. Engineers seek your feature engineering expertise. Product trusts your model interpretations. Your research directly shapes roadmap decisions. Your Arena You'll own data science for one of our core intelligence systems: Predictive SEO Intelligence - Models that forecast rankings, predict algorithm impacts, and identify optimization opportunities before competitors. Agent Decision Systems - ML powering autonomous agent behavior. Reward modeling, multi-armed bandits, reinforcement learning from human feedback. Content & Entity Intelligence - NLP systems for semantic analysis, entity extraction, content quality scoring, and generative optimization.

Requirements

You've shipped ML systems that drive real business outcomes - and you can prove it.
7+ years applied data science, with 3+ years in production ML environments (not just research).
Deep Python fluency - NumPy, Pandas, Scikit-learn, PyTorch or TensorFlow. You write production code, not just prototype scripts.
SQL at scale - complex queries against terabyte datasets, query optimization, ClickHouse or analytical databases.
ML engineering mindset - feature stores, model serving, A/B testing, monitoring, drift detection. You think in pipelines, not just algorithms.
AI-native workflow - you use AI coding tools to accelerate research, automate analysis, and build faster.
Scientific communication - you explain complex methodologies to executives with clarity and confidence. No jargon, just impact.
Research leadership - you've mentored others, defined research agendas, and shipped models that required cross-functional coordination.

Nice To Haves

NLP/LLM expertise.
Search or SEO domain knowledge.
Reinforcement learning or bandit algorithms.
Published research or active ML community presence (Kaggle, GitHub, papers).

Responsibilities

Ship Production ML
Build and deploy models that serve predictions in milliseconds. Your work runs in production, not in notebooks.
Own the full ML lifecycle: feature engineering, training pipelines, model validation, A/B testing, and monitoring for drift.
Design experiments that prove business impact, not just statistical significance. You measure revenue lift, not just accuracy.
Architect Intelligence
Define our research roadmap. What should we predict? What should we optimize? Your scientific judgment sets priorities.
Design feature pipelines that handle terabyte-scale data. ClickHouse, PostgreSQL, streaming signals - you architect data flows for model performance.
Build evaluation frameworks that catch model degradation before customers do. You own model reliability.
Accelerate with AI
Leverage AI coding tools (Claude Code, Copilot, custom agents) to 10x your research velocity. You automate the repetitive, focus on the breakthrough.
Build internal tools that democratize data science across the company. Your abstractions make others faster.
Lead Through Science
Mentor data scientists and ML engineers through code review, research review, and technical guidance. You elevate the team's scientific rigor.
Translate complex methodologies into clear business impact. You present to leadership with confidence and craft.
Partner with Engineering, Product, and AI teams to integrate models into production systems. You don't hand off research; you co-own deployment.