Role Overview
Own the end-to-end software and firmware architecture for advanced robotic platforms — spanning humanoid robots and autonomous drones (UAVs) — built on edge-AI SoMs from NVIDIA (Jetson AGX Orin, Orin NX, Orin Nano), Qualcomm (Dragonwing), and MediaTek (Genio). You will define the compute architecture, real-time control stack, perception/AI pipeline, and autonomy framework, working closely with the electronics, mechanical, and mechatronics teams to translate subsystem designs into cohesive, safe, and deployable software across platforms and silicon vendors.
Key Responsibilities
- Define the overall software architecture spanning low-level firmware, real-time control, middleware, autonomy, and high-level AI/perception layers, reusable across humanoid and UAV platforms.
- Architect the compute strategy across heterogeneous SoMs — NVIDIA Jetson, Qualcomm Dragonwing, and MediaTek Genio — partitioning workloads across CPU, GPU, NPU/DLA, DSP, and companion MCUs/flight controllers.
- Design real-time control frameworks for both domains: actuator/joint, balance, and locomotion control for humanoids; flight control, stabilization, and navigation for drones.
- Establish deterministic communication with motor controllers, ESCs, and flight controllers (EtherCAT, CAN/CAN-FD, SPI, UART, PX4/MAVLink, DroneCAN/UAVCAN).
- Define the robotics middleware and framework strategy (ROS 2, DDS, PX4/ArduPilot integration, real-time Linux / PREEMPT_RT).
- Architect the perception and sensor-fusion pipeline: multi-camera vision, depth/stereo, LiDAR, radar, IMU, GPS/GNSS, and force/torque sensing, including calibration and time synchronization.
- Define the autonomy stack: state estimation, SLAM / visual-inertial odometry, localization, mapping, path/motion planning, obstacle avoidance, and behavior/mission management.
- Lead integration of vendor AI stacks — NVIDIA (Isaac ROS, CUDA, TensorRT, DeepStream), Qualcomm (QNN / AI Engine, Neural Processing SDK, QAIRT), and MediaTek (NeuroPilot, Genio SDK) — and optimize model deployment (quantization, pruning, on-device inference).
- Own the board bring-up and BSP strategy across vendors — NVIDIA L4T/JetPack, Qualcomm/MediaTek Yocto and Linux/Android BSPs, device tree, kernel drivers, bootloaders, boot and flashing workflows — with the hardware team.
- Define the safety architecture: fail-safe and emergency-stop behavior, redundancy, watchdogs, return-to-home/geofencing for UAVs, fall detection and safe-state handling for humanoids.
- Establish the power, thermal, and resource-management strategy for battery-powered, weight- and thermally-constrained platforms.
- Define connectivity and telemetry: Wi-Fi, 5G/LTE, RF links, ground control station / operator interfaces, and secure remote command and monitoring.
- Own OTA update, device provisioning, security (secure boot, encryption, key management), and data-logging/black-box strategy.
- Establish the simulation and digital-twin strategy (Isaac Sim, Gazebo, hardware-in-the-loop, software-in-the-loop) and sim-to-real transfer.
- Define coding standards, architecture documentation, CI/CD, automated testing, and the overall development and release process.
- Mentor junior engineers and provide technical direction across the software team; support hiring and technology evaluation.
Required Skills & Experience
- 4+ years in embedded software/firmware, with demonstrated architecture ownership on complex real-time systems.
- Hands-on experience with edge-AI SoMs — NVIDIA Jetson and at least one of Qualcomm Dragonwing or MediaTek Genio — including bring-up, BSP, and workload optimization across GPU/NPU/DSP.
- Strong C/C++ for real-time and embedded systems; Python for tooling, scripting, and prototyping.
- Deep understanding of real-time Linux (PREEMPT_RT), RTOS concepts, Yocto-based BSPs, bootloaders, and deterministic control loops.
- Robotics middleware expertise, especially ROS 2 and DDS; experience designing modular, distributed node architectures.
- Experience with motor control, ESCs, servo/actuator drives, and industrial/robotics buses (EtherCAT, CAN-FD, UART, SPI, I2C).
- Familiarity with flight stacks and protocols (PX4, ArduPilot, MAVLink, DroneCAN) and/or humanoid whole-body control.
- Perception and sensor fusion: multi-sensor integration, calibration, time sync, SLAM, visual-inertial odometry, and Kalman/EKF-based state estimation.
- GPU/NPU-accelerated AI across vendor toolchains: CUDA/TensorRT, Qualcomm Neural Processing SDK/QNN, MediaTek NeuroPilot; model optimization and on-device inference.
- Understanding of control theory, kinematics/dynamics for high-DOF systems, and/or flight dynamics and control.
- Experience with functional safety, redundancy, and fault-tolerant design for autonomous or safety-critical systems.
- Familiarity with device security: secure boot, encrypted storage, OTA, and secure communications.
Nice to Have
- Experience porting or benchmarking AI workloads across multiple SoC vendors (NVIDIA, Qualcomm, MediaTek).
- Production experience with humanoid, legged, or high-DOF robots and/or UAV/drone autonomy at scale.
- NVIDIA Isaac Sim / Isaac ROS, Omniverse, Gazebo, or other sim-to-real and reinforcement-learning workflows.
- Knowledge of whole-body control, MPC, reinforcement learning for locomotion, or autonomous navigation and path planning.
- Experience with regulatory/compliance frameworks (aviation/UAV regulations such as DGCA/FAA/EASA, functional-safety standards).
- Exposure to fleet management, multi-robot / swarm coordination, and cloud-edge architectures.
2. Junior Software Design Engineer — Robotics & Autonomous Systems
Role Overview
Support the development of the software and firmware stack for humanoid robots and autonomous drones on edge-AI SoMs from NVIDIA (Jetson), Qualcomm (Dragonwing), and MediaTek (Genio). You will implement, test, and integrate software modules under the guidance of the Software Architect and senior engineers, growing into real-time control, perception, autonomy, and embedded systems across platforms and silicon vendors.
Key Responsibilities
- Implement and test software modules across firmware, control, perception, and application layers for humanoid and/or UAV platforms.
- Assist with SoM bring-up tasks: flashing/board setup on Jetson (JetPack/L4T), Qualcomm Dragonwing, and MediaTek Genio; configuring peripherals and validating drivers.
- Develop and maintain ROS 2 nodes for sensing, actuation, navigation, and data handling.
- Write interface and driver code for sensors, actuators, ESCs, and flight controllers over CAN, SPI, I2C, UART, USB, and MAVLink.
- Support integration of cameras, LiDAR, IMU, and GPS/GNSS, including data capture, calibration, and logging.
- Assist with deploying and testing AI/ML models on-device using vendor toolchains (TensorRT, Neural Processing SDK, NeuroPilot).
- Build test scripts, telemetry, logging, and diagnostic tools; support integration, bring-up, and debugging on hardware and in simulation.
- Reproduce and help resolve issues found in bench testing, field trials, and flight/motion tests.
- Contribute to documentation of code, interfaces, and test procedures; participate in code reviews.
Required Skills & Experience
- 1–3 years of experience (or strong internship/academic-project background) in embedded software or robotics.
- Proficiency in C/C++ and Python.
- Familiarity with Linux development environments, Git, and basic shell scripting.
- Exposure to ROS/ROS 2 and embedded communication protocols (CAN, I2C, SPI, UART).
- Understanding of microcontrollers, sensors, actuators, and basic real-time concepts.
- Basic understanding of control loops, state estimation, or computer-vision fundamentals.
- Eagerness to learn GPU/NPU-accelerated computing, flight control, and robotics motion control.
- Good debugging skills and a hands-on, hardware-friendly attitude.
Nice to Have
- Hands-on experience with NVIDIA Jetson, Qualcomm Dragonwing, MediaTek Genio, or other SoM/SBC platforms.
- Exposure to vendor AI toolkits (TensorRT, Qualcomm Neural Processing SDK, MediaTek NeuroPilot) or ML frameworks (PyTorch, TensorFlow).
- Exposure to drone flight stacks (PX4, ArduPilot, MAVLink) or robotics control.
- Familiarity with simulation tools (Gazebo, Isaac Sim) and hardware-in-the-loop testing.
- Academic or hobby projects in robotics, drones, motion control, or computer vision.
- Familiarity with CI/CD and unit testing.
Pay: ₹500,000.00 - ₹600,000.00 per year
Benefits:
Work Location: In person
