Data Scientist - Data-as-a-Service

📋 Description

• • Own the full AI data lifecycle for Fortune-500 and cutting-edge research partners: from raw, messy enterprise data to production-grade, model-ready datasets that power mission-critical AI systems.
• • Architect and implement end-to-end ML pipelines that combine programmatic labeling, weak supervision, active learning, and human-in-the-loop review to turn subject-matter-expert knowledge into high-quality training data at unprecedented speed and scale.
• • Translate ambiguous business objectives into crisp technical specifications, then deliver against them by writing production Python, SQL, and PySpark code that runs on petabyte-scale data lakes and cloud-native infrastructure (AWS, GCP, Azure).
• • Design statistically rigorous evaluation frameworks—precision-recall curves, confidence intervals, error analysis dashboards—that give executives and ML engineers the evidence they need to trust the data and the models built on top of it.
• • Build reusable, Snorkel-internal libraries and micro-services that accelerate future engagements: think auto-tuned labeling functions, active-learning samplers, or real-time data-quality monitors that cut weeks of manual work down to hours.
• • Collaborate daily with a cross-functional squad of data engineers, ML researchers, product managers, and customer SMEs to unblock bottlenecks, prioritize features, and ship incremental value every sprint.
• • Present findings and recommendations to C-suite stakeholders and technical leads through clear slide decks, live notebooks, and interactive dashboards that turn complex data stories into actionable next steps.
• • Continuously improve our human-in-the-loop processes by running A/B tests on reviewer instructions, UI/UX tweaks, and ML-assisted pre-labeling to maximize throughput without sacrificing quality.
• • Contribute to Snorkel’s open-source ecosystem and research publications, sharing novel techniques in weak supervision, data-centric AI, and evaluation methodology with the broader community.
• • Travel up to 20 % for quarterly on-site workshops with strategic customers, where you’ll whiteboard solutions, mentor client data teams, and gather feedback that directly shapes our product roadmap.
• • Stay on the bleeding edge of generative-AI and LLM tooling, experimenting with prompt engineering, retrieval-augmented generation, and synthetic data techniques to keep our delivery toolkit state-of-the-art.
• • Champion a culture of reproducibility and engineering excellence: peer-review PRs, enforce test coverage, instrument observability, and document tribal knowledge so every project is a stepping-stone—not a one-off.