As an Android Software Engineer, you'll lead the development of a mobile-first AI platform that engages users in everyday contexts. Your work will center on crafting a robust, efficient Android client that delivers seamless interactions across chat, vision, voice, and personalized recommendations.
What You'll Do
- Develop and maintain production-grade Android applications using Kotlin, with a focus on stability and responsiveness.
- Integrate AI-driven features by connecting to backend services and managing asynchronous data flows.
- Design intuitive interaction patterns for streaming outputs, handling retries, and displaying partial results.
- Optimize app performance, memory footprint, and latency, especially during intensive AI workflows.
- Implement telemetry, logging, and user feedback systems to inform AI model improvements.
- Work closely with backend and machine learning teams to define API behavior and ensure consistent system performance.
- Ensure the app meets high standards for security, scalability, and resilience in unpredictable network conditions.
What We're Looking For
- At least three years of professional Android development with Kotlin.
- Proven experience integrating AI features such as large language models, computer vision, or speech processing.
- Strong grasp of asynchronous programming using Coroutines and Flow.
- Familiarity with REST or gRPC APIs and data serialization formats.
- Solid skills in debugging, profiling, and diagnosing performance bottlenecks.
- Experience with on-device inference frameworks like MLKit or TensorFlow Lite.
- A published app on the Google Play Store demonstrating production-quality development.
Technology Environment
You'll work primarily with Kotlin and Java, leverage SQL and noSQL data stores, and implement on-device AI models using TensorFlow Lite.
Team & Culture
You'll join a high-performing, hands-on team that values technical excellence, rapid iteration, and shared decision-making. We value individuals who can work independently, bring clarity to complex problems, and balance fast execution with continuous learning.
