Lead Embedded Software Engineer

We are seeking a Lead Embedded Software Engineer to lead software and AI exploration and development for AI-embedded smart hardware and software, keyboards, mice, gaming peripherals, and IoT devices.

You will own the software stack that makes hardware intelligent: on-device ML for personalization and predictive input, LLM-backed companion app features, and natural language device control across edge, app, and cloud.

Key Responsibilities:

AI Agent Integration & Feature Development

• Develop on-device AI features for smart peripherals (input personalization, predictive shortcuts, gesture/intent classification, adaptive behavior driven by usage patterns)

• Build LLM-backed companion app features (natural language device configuration, AI assistant integration, conversational automation and settings control)

• Integrate AI agent frameworks into the product stack to deliver and orchestrate smart hardware features across device, companion app, and cloud

• Integrate the AI agent with chat/messaging platforms (e.g., Feishu, WeChat, Telegram, Discord) for remote device control, status notifications, and event-driven automation

Device Management & Cloud Platform

• Design and implement the device management backend: provisioning, settings sync, OTA distribution, and companion app APIs

• Build the telemetry pipeline: ingest sensor/usage events from devices, feeding product analytics and AI model improvement loops

• Integrate China-local LLM and AI service providers for companion app AI features; implement provider abstraction with config-driven failover and secure API key management

• Achieve <.5s end-to-end response latency from user request to AI feature delivery, using streaming APIs and edge-optimized provider routing

LLM & AI Provider Integration

• Configure the multi-provider LLM architecture for China-local providers as primary (MiniMax, Qwen, Kimi, DeepSeek, Zhipu AI) supporting companion app AI features and natural language device control

• International providers (Claude, GPT) as fallback

• Implement provider switching via config (no code change), secure API key management, and graceful fallback when primary LLM is unavailable

Sensor Integration & Signal Processing

• Integrate optical sensors, hall-effect/TMR sensors, IMU, haptic actuators, or microphones into a software-controlled AI feature pipeline (e.g., adaptive input, gesture recognition, ambient awareness)

• Expose sensor data streams to on-device AI models and companion app cloud services for real-time inference

• Collaborate with HW/EE engineer on sensor calibration, signal conditioning, and device protocol data formats

Hardware-Software Interface

• Work closely with EE/firmware engineer to define the USB/BLE/I2C/UART communication protocol between the peripheral MCU and the host-side software stack

• Implement the host-side Hardware Controller layer: translate high-level AI feature commands (e.g., “set haptic pattern X”, “trigger RGB effect”) into device protocol messages

• Implement real-time device state feedback (latency, battery, sensor readings) consumed by the companion app and AI agent for adaptive behavior

• Implement physical feedback behaviors (haptic patterns, audio cues, LED indicators) triggered by AI agent decisions or user notifications

Required Qualifications

Technical Skills

• Programming Languages: Proficient in Node.js/JavaScript and Python; comfortable with C or C++ (for reading protocol specs and writing lightweight host-side native modules)

• AI Agent Platforms: Hands-on experience with AI agent frameworks (e.g., LangChain, LangGraph, Pipecat, or similar), including creating custom skills/plugins for product feature delivery

• Embedded Linux: Experience developing software on ARM-based SBCs (Raspberry Pi, Orange Pi, Jetson Orin Nano, etc.)

• LLM APIs: Practical experience integrating LLM provider APIs , especially China-local providers (MiniMax, Qwen, Kimi, DeepSeek)

• Backend Development: REST/WebSocket APIs for companion apps, device provisioning, and cloud integration for AI-powered smart hardware features

• Sensor & Signal Processing: Basic experience with sensor data pipelines (IMU fusion, optical flow, audio signal processing, or OpenCV/MediaPipe)

• Linux Systems: Comfortable with Linux system administration, systemd services, udev rules, USB/HID debugging, and profiling on resource-constrained ARM hardware

Experience

• 5+ years of software development experience

• 1+ year working with Linux application development (system services, daemons, IPC, process management)

• 1+ year working with LLM-based applications or AI agent systems

• Demonstrated experience with at least one smart hardware, IoT, or physical computing project

• Experience with real-time signal processing on embedded platforms

Preferred Qualifications

• Experience with Dify, LangGraph, or MCP (Model Context Protocol) for building tool-integrated agent pipelines

• Experience with chat platform API integration (e.g., WeChat, Feishu, Telegram, Discord)

• Knowledge of HID protocols (USB HID, BLE HID/HOGP) and peripheral MCU communication (UART, SPI, I2C, USB CDC-ACM)

• Experience with embedded display integration (LVGL, SDL2, or framebuffer) for smart peripheral UIs

• Experience building web-based companion app configuration UIs (device settings, OTA trigger, provisioning)

• Experience with OTA firmware distribution (staged rollouts, A/B cohorting, rollback)

• Chinese language proficiency (for coordinating with China-based vendors and LLM providers)

• Prior experience with AI robot, voice companion, or social robot products is a plus