We’re looking for an Android Applications Developer to design, build, and continuously evolve production-grade Android/AAOS applications and services.
You will focus on integrating prebuilt ML models into in-vehicle environments, developing and maintaining end-to-end pipelines (capture → preprocessing → inference → UX/telemetry), and ensuring scalability, reliability, and performance over time.
This role supports not only immediate deliverables but also a broader framework for future ML- powered applications across our platforms.
Responsibilities:
● Integrate prebuilt ML models (e.g., TFLite, MediaPipe, ONNX Runtime) into Android/AAOS apps and services with clear and maintainable model I/O contracts.
● Offline-first inference; optional cloud inference path (REST/gRPC) using the hosted model endpoint; implement routing and fallback between local and cloud paths.
● Develop robust, real-time audio and sensor capture pipelines (AudioRecord/AAOS audio APIs; optional car audio plugin service) with buffering and back-pressure handling.
● Implement model pre/post-processing as specified by the ML team (e.g., windowing, normalization, log-mel) using provided reference code and evolving best practices.
● Design and maintain event logic (thresholds, debouncing, hysteresis) and configuration toggles; collaborate with ML and Product to calibrate and adapt over time.
● Optimize apps for latency, memory, and power efficiency; select runtime delegates (NNAPI/GPU/DSP) when appropriate; profile and tune cold-start and steady-state performance.
● Build developer-facing tools and lightweight UIs (Jetpack Compose) for debugging, telemetry visualization, tracing, threshold management, and runtime selection.
● Implement privacy-preserving telemetry and evaluation hooks (e.g., precision/recall estimates, false positive rates) without retaining raw audio or sensitive data.
● Establish quality gates, including unit/instrumentation tests, Compose UI tests, and contract tests that validate model interfaces, shapes, and versioning.
● Document architecture decisions, risks, and integration learnings; contribute to productionization strategies and ongoing platform improvements.
● Collaborate across multidisciplinary teams to ensure smooth deployment, maintainability, and scalability of ML features in-vehicle.
Skills and Qualifications Required:
● Production-level Android development in Kotlin, with expertise in coroutines/Flows, dependency injection, background execution, and modern architecture patterns (MVVM/MVI).
● Proven experience integrating ML models on Android (TFLite/MediaPipe/ONNX Runtime) and/or invoking cloud models from mobile apps with secure auth, retries, and timeouts.
● Strong understanding of Android audio capture (AudioRecord), streaming pipelines, and latency-aware processing.
● Performance engineering skills, including profiling with Perfetto/Traceur/Android Studio Profiler; startup, jank, and memory optimization.
● Experience with testing frameworks (JUnit, instrumented tests, Compose testing) and interface/contract testing for model boundaries.
● Effective collaboration and documentation skills suited to fast-moving development cycles.
● Strong problem-solving mindset with attention to detail, clarity in communicating trade-offs, and the ability to operate autonomously.
● Familiarity with GitHub Pull Request and code review processes.
Preferred:
● AAOS or embedded Android experience; knowledge of in-vehicle UX and system constraints.
● Android car audio plugin service experience.
● Background in experiment frameworks, analytics, or safety-relevant alerting.
● Understanding of privacy-preserving telemetry and compliance requirements.
● Familiarity with the Android Neural Network API (NNAPI).
● Experience with Qualcomm’s AI Hub and AI Runtime SDKs.
Benefits:
● Family health plan.
● Birthday day off.
● Continuous training through content platforms.
And more!
Perfecto, seguí hablando con él.
Aplica para esta posición
Si ya estás hablando con un reclutador de CONEXIONHR, NO COMPLETES EL FORMULARIO.
