Senior Data Engineer, Synthesia

• You're an experienced Data Engineer who loves building robust, reproducible, and scalable data systems and you care deeply about data quality, performance, and usability
• A track record of owning data pipelines end-to-end - from data sourcing and transformation through to modeling and observability
• Experience designing with data architecture patterns like lambda or kappa, and strong opinions on ETL vs ELT
• Familiarity with orchestration patterns (DAG-based, stateless, etc.) and workflow tools (we use Kubeflow and Spark)
• Strong SQL skills and experience building data models using patterns like dimensional or medallion architecture
• Hands-on experience with columnar formats and open table standards (Delta Lake, Iceberg, Hudi)
• Proficiency in Python and infrastructure tools like Terraform. Experience hosting open source frameworks
• Most importantly, you start with the problem & not the tech. You build data systems based on real use cases and are thoughtful about tradeoffs, complexity, and cost

Desirable

• The Data team powers Synthesia’s R&D efforts by providing scalable, reliable, and observable data infrastructure for our Applied Research team. The Data team is responsible for all data needs, from ingestion to transformation, to modeling
• As a Data Engineer, you’ll take full ownership of pipelines that move and structure massive volumes of non-tabular video, audio, image, and text data across our lakehouse architecture
• You’ll work on high-throughput systems that enable internal teams to query and consume data more efficiently, and you’ll help drive the architecture, implementation, and observability of our entire data platform
• This is not a traditional data engineering role. It’s hands-on, hybrid, and high-impact, sitting at the intersection of research, infrastructure, and product.
• You’ll partner closely with R&D and infra teams, helping ensure that our systems scale with speed, performance, and cost in mind.
• We’re looking for someone who starts from the problem, not the tooling and knows how to build data systems around real-world use cases
• Re-architecting our ingestion pipeline to better support long-form, high-resolution video data
• Implementing data tiering and modeling patterns like medallion architecture or dimensional modeling
• Host and manage open-source tools and frameworks
• Designing scalable observability for data quality and pipeline health
• Supporting researchers by enabling faster, more targeted data retrieval at training time