Data Scientist- Machine Learning & AI

Who we are Lightricks is an AI-first company bridging imagination and creation through innovative tools for content creation. Our flagship app, Facetune, is used by millions of creators worldwide for AI-powered photo and video editing. Combining advanced computational graphics and generative AI, we empower creators to push the boundaries of what's possible. Our Data Science team sits at the center of how Facetune makes business decisions. We're a small team with a direct line to leadership — our models are the actual inputs to revenue planning, marketing efficiency (budget allocation, campaign optimization), and product strategy. Our work has led to published research — we go deep when the problem calls for it. The role We're hiring a strong Data Scientist with deep understanding in classical ML skills and experience with unstructured data. You'll build prediction and optimization models across the business, and help us expand into new territory: extracting signals from user prompts, photos, and video to power ad prediction, user intent modeling, recommendations, and personalization. You'll also contribute to the team's AI ecosystem, including evaluating and calibrating AI agent outputs. Our stack: Python, BigQuery, GCP, our own production pipeline framework (Nova), Prefect, and Grafana. We're looking for someone with strong modeling instincts who cares about business impact — not just technical depth, but making sure the work lands and changes how decisions are made. What you will be doing Build prediction, classification, optimization, and causal inference models — revenue forecasting, campaign performance prediction, user segmentation, spend allocation, personalization, and other ML projects across the team's portfolio Design and maintain models in both batch and online serving environments — owning the full path from research to production Evaluate, calibrate, and improve AI agent outputs — designing benchmarks, measuring accuracy, and identifying failure modes in AI-generated results Research and adopt new models and AI technologies to expand what the team can deliver — finding practical ways to scale our impact beyond what we can build manually Work with rich, large-scale user data — but the challenges here are deep: building models that stay accurate as the product and market shift, that scale reliably, and that keep driving decisions over time, not just at launch Extract signal from unstructured user data — prompts, photos, ad creatives, video interactions — using LLMs, embeddings, or custom approaches where domain-specific signal extraction matters Lead projects end-to-end: from identifying a business problem with stakeholders, through research and planning, to production deployment and monitoring Work directly with Finance, Marketing, and Product stakeholders to define problems, align on approach, and deliver results they act on Your skills and experience M.Sc. or Ph.D. in Statistics, Computer Science, Mathematics, or a related quantitative field 5+ years of hands-on data science experience Strong foundation in classical ML — classification, regression, and time-series forecasting Experience working with unstructured data (text, image, or video) alongside structured data, in a production or applied research setting Strong programming skills in Python and SQL (BigQuery or equivalent) Experience with ML frameworks such as scikit-learn, XGBoost, or similar; familiarity with time-series methods (Prophet, ARIMA) is a plus End-to-end project experience: from business problem definition through research, production, and stakeholder delivery Daily use of AI tools in your workflow — you already work with LLMs, coding assistants, or AI-powered tooling as a natural part of how you get things done Strong communication skills — ability to present complex technical findings clearly to non-technical audiences, write structured documentation, and collaborate effectively across teams Good to haves Experience with NLP, computer vision, or multimodal feature extraction in applied settings — including integrating LLMs into production models Background in marketing analytics, revenue modeling, or other business-facing DS problems Experience evaluating, calibrating, or benchmarking LLM and AI agent outputs Experience with anomaly detection or unsupervised methods Consumer app or B2C product experience