• Maintain Performance - Monitor the architecture and identify opportunities to optimize code to maintain and enhance mainframe performance. • Debug and Test - Debug, troubleshoot, and analyze complex system interactions to keep mission-critical applications running. • Develop for the Future - Extract business logic from legacy systems for migration or integration with cloud-based services.

• Build and ship full-stack features across Deepgram’s customer-facing Console and Playground experiences. • Build internal-facing tools and services that improve workflows and productivity across Deepgram. • Design, build, and maintain APIs and backend services that power end-user and internal product experiences. • Partner with product, design, and engineering to create accessible, performant, and responsive interfaces backed by reliable systems. • Translate product ideas and requirements into well-scoped, incrementally releasable work across frontend and backend layers. • Balance rapid iteration with long-term maintainability, quality, scalability, and performance. • Help raise the bar through code reviews, technical design discussions, and strong engineering practices across the stack.

• Contribute to the technical direction of the Perception Architecture team and the broader perception organization. • Design and develop core machine learning components and architectures that enable the perception system to scale and reliably address complex, long-tail driving scenarios. • Drive the initial implementation and prototyping of key architectural decisions, such as handling long tail challenges or long range detection. • Contribute to the improvement of model frameworks, architecture, and data pipelines to eliminate development bottlenecks and accelerate model iteration. • Lead challenging technical topics and build consensus across sister teams (e.g., Actor Detection, Scene Understanding, Tracking) and stakeholders to ensure autonomy-wide alignment.

• Embed with a team. Work alongside a function for sustained periods, joining their day-to-day so you understand how they really operate. • Map how the work gets done. Build a clear picture of the function's current workflows and surface where time, handoffs, rework, and manual effort pile up. • Find where AI creates the most value. Prioritize the opportunities that will move the function's outcomes the most, and align with the team on what to tackle. • Build it in production. Ship real, AI-powered workflows that solve the problems you find, working hands-on with the team and our platform. • Scale what works. Turn proven solutions into reusable capabilities so a win with one team accelerates the whole organization. • Own the outcome. Measure success by the business results your partners achieve: revenue generated, cost saved, cycle time reduced, manual work eliminated.